Meeting of the Joint Subcommittee on Reliability and Probabilistic Risk Assessment, Plant Operations, and Regulatory Policies and Practices
UNITED STATES OF AMERICA
NUCLEAR REGULATORY COMMISSION
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS
MEETING: RELIABILITY AND PROBABILISTIC RISK
ASSESSMENT, PLANT OPERATIONS, AND REGULATORY
POLICIES AND PRACTICES
11545 Rockville Pike
Tuesday, July 13, 1999
The Subcommittees met, pursuant to notice, at 1:00 p.m.
GEORGE APOSTOLAKIS, Chairman of the Subcommittee
on Reliability and PRA
THOMAS KRESS, Chairman, of the Subcommittee on
Regulatory Policies and Practices
JOHN BARTON, Member, ACRS
ROBERT SEALE, Member, ACRS
ROBERT UHRG, Member, ACRS
GRAHAM WALLIS, Member, ACRS
MARIO BONACA, Member, ACRS
P R O C E E D I N G S
DR. APOSTOLAKIS: The meeting will now come to order. This
is a meeting of the ACRS Subcommittees on Reliability and Probabilistic
Risk Assessment and on Regulatory Policies and Practices.
I'm George Apostolakis, Chairman of the Subcommittee on
Reliability and PRA. Dr. Kress is the Chairman of the Subcommittee on
Regulatory Policies and Practices.
ACRS members in attendance are John Barton, Robert Seale,
Bob Uhrig, Graham Wallis, Mario Bonaca.
The purpose of this meeting is to review proposed options
for development of risk-informed revisions to 10 CFR Part 50, domestic
licensing of production and utilization facilities.
The subcommittees will gather information, analyze relevant
issues and facts, and formulate proposed positions and actions as
appropriate for deliberation by the full committee.
Dr. Powers just joined us.
Michael T. Markley is the Cognizant ACRS Staff Engineer for
The rules for participation in today's meeting have been
announced as part of the notice of this meeting previously published in
the Federal Register on June 21, 1999.
A transcript of the meeting is being kept and will be made
available as stated in the Federal Register notice.
It is requested that speakers first identify themselves and
speak with sufficient clarity and volume so that they can be readily
We have received no written comments or requests for time to
make oral statements from members of the public.
I forgot to mention Mr. John Sieber, who is also
participating, although not quite a member yet, I understand, as an
The ACRS reviewed SECY-98-300 in December 1998 and issued a
report to the Commission, dated December 14, 1998. In a staff
requirements memorandum, dated June 8, '99, the Commission directed the
staff to pursue option two, develop risk-informed definitions for
safety-related and important to safety systems, structures and
components; to modify the scope of the maintenance rule; approved option
three to study other areas of 10 CFR Part 50 that may be risk-informed;
and, approved policy issues related to voluntary provisions of the use
of exemptions for pilot plants.
We will hear about the staff's plans today for risk-informed
on these issues.
The staff also met with NEI on June 22 and July 8 of this
year on this matter. NEI is supposed to go first, according to the
schedule, but NEI is not here.
So we shall proceed with the staff's presentation, which was
scheduled for 1:45. So we're ahead of the schedule already.
Who will take the lead and introduce everybody? Okay.
MR. NEWBERRY: I'll take care of that, Mr. Chairman. My
name is Scott Newberry. I'm the Deputy Director in the Division of
Regulatory Improvements in NRR. Let me introduce the folks with me. At
either end, Gary Holahan and Tom King, who you've seen before on
risk-informed activities. Of course, Mark Cunningham over here in front
of the projector, and to my right, a new addition to our group, Tim
Reed, who is spent the last few years struggling with steam generator
issues, but now we've drawn him into the risk informing Part 50
Later in the presentation, there is a specific viewgraph on
the organization in the NRC working this project in terms of the role of
NRR, in terms of the role of Research. I won't go into that in any
detail here in the introductory remarks, only to say that this is a
large project. It's going to touch on the programs in every division in
NRR and we're working closely and have established a process that will
talk to you about, working with the Office of Research on both options,
and we'll try to make that clear to you and answer your questions.
But there's certainly a larger group working on this project
in some of the other activities because of its breadth and touching on
many regulations, some of which we will talk to you about today.
We're flexible on the agenda. We'll just plan on going
ahead and proceeding, Mr. Chairman, and if you want to make a change,
DR. APOSTOLAKIS: Let's see what -- go ahead.
MR. NEWBERRY: Okay. Let's go ahead. Mark?
MR. CUNNINGHAM: Thanks, Scott. There are six parts to the
presentation today. The first couple I'm going to cover, talking about
what was talked about in SECY-98-300 and what the Commission told us to
do in the SRM. I will talk a little bit about the organization of the
After that, we're going to talk about the plans for option
two. Tim will cover that. We'll come back then to some initial ideas
and plans for option there that I will talk about. And last, but not
least, we've got some technical process and policy issues and some
schedule, and Tim King will talk about those.
There were three key elements to SECY-98-300. We started in
the document by describing what we saw as the intended outcome of
modifications to Part 50; also talked about three options for
accomplishing that outcome; and then we raised four policy issues that
went with those. I'm going to come back and talk about each of these in
a little more detail.
Now, as laid out in 98-300, the intended outcome of a
modification to Part 50 was to have a Part 50 or some new version which
accomplishes the following at the end. One, it would do, as Part 50
does today, to provide reasonable assurance of adequate protection of
the public health and safety.
Second, it would --
DR. POWERS: When you say that this new Part 50, for want of
a better term for it, would do as Part 50 does today, to provide
reasonable assurance of adequate protection of the public health and
safety, do you mean exactly that way, in a very rather nebulous
definition of what reasonable assurance is?
MR. CUNNINGHAM: Not necessarily.
DR. POWERS: You would be looking for something that is
crisp, comprehensible, explicit.
MR. KING: We're not at the point of recommending a
definition of adequate protection. We're farming that out.
DR. POWERS: It was more that you would have a definition
and giving me the definition. So your intention is to actually define
what adequate protection means.
MR. KING: At this point, we haven't made that decision.
DR. WALLIS: Let me follow up on that. Will the public
health and safety be better or worse as a result of what you've done?
MR. CUNNINGHAM: In both cases, we expect that it will be,
as the words say, that we will have reasonable assurance that it is
DR. WALLIS: That's not good enough. Unless you have a
metric or something to compare it, you have no idea if the new one is
any better than the old one, and reasonable assurance doesn't mean
anything to me.
MR. NEWBERRY: Let me try. I don't know that we're going to
DR. WALLIS: If you're going to be quantitative about risk,
you've got to be quantitative about number one, as well.
MR. NEWBERRY: We're embarking on a rulemaking plan in
process here, as we would with any rulemaking, and going into it, we
don't have an answer to your question with respect to making things
better or worse.
We'll get into the context of the discussion when we talk
about the option in terms of a risk-informed scope, but we're going to
have to have an answer to your question as we proceed into the process
DR. WALLIS: Unless you have answers which are quantitative,
you are just invoking words.
MR. KING: We're going to hit this when we get to the policy
issues. One of the policy issues is do we take the current risk profile
and make small changes around it so that each plant, which has a
different risk profile, is just adjusting around where it currently is,
or do we take some more absolute approach and try and come up, for
example, with a Part 50 that brings everybody down to a level of safety
of the safety goals.
We haven't decided one way or the other at this point. I
think the Commission ultimately has to make that decision. It's one of
the issues we're going to talk about as we get into the presentation.
DR. SEALE: Is the intent, though, to have a quantitative
metric, if you will, on the revised definition of adequate protection of
the public health and safety?
MR. CUNNINGHAM: That is not -- it's not clear at all that
that would be the outcome of this.
DR. SEALE: My point being that to come up with a
quantitative measure of what the present words mean now really doesn't
necessarily help you one way or the other. You could very easily say
I'm just going to pass that by and go for a quantitative version of what
it will mean in the future.
And while that may not be a this versus that comparison, as
Dr. Wallis has urged, at least it gets us on a quantitative basis.
MR. HOLAHAN: This is Gary Holahan. Let me go back a little
bit here. What we're talking about is risk-informing Part 50, not risk
basing Part 50. We've talked for a long time with this committee and a
lot of other people in coming to a consensus on how to use not only
quantitative risk assessment, but to bring other valuable safety
concepts, like defense-in-depth and safety margins, which are sometimes
quantitative, but frequently qualitative, into the decision-making
And I don't think that we want to be or that we are on the
course to converting all of our safety decisions into numerical
DR. WALLIS: I think you owe it to the public to tell them
how safe they are and not say we are risk-informed; therefore, whatever
we do is adequate protection.
MR. HOLAHAN: That's fair, that's fair.
MR. KING: We're taking the approach that whatever we do
defines adequate protection. The current regulations provide some level
of safety above adequate protection. That's been the agency's position
for a long time.
I would expect the risk-informed set would do the same
thing; do we have to define adequate protection to quantify that? It's
not clear to me we do.
And you're asking questions that we don't have answers to
because what we're talking about today is work in progress. This is a
status report. This is kicking around some ideas, some options and so
forth. We're not here to present a final proposal.
DR. WALLIS: I guess what you're hearing, though, is that
there are several members on this side of the table who would like you
to be more quantitative and measurable about these broad statements.
Then we have something more definite to get hold of.
Work in progress doesn't mean too much -- to put it a
different way, work in progress is more meaningful if you put it in a
context of metrics. You have something to go by, some telling what's
been achieved and what you're aiming at, rather than some philosophical
statement, some definite goal which we can then check and see if you've
met it or not.
I think we ought to move on, but I think you're seeing the
way that -- you're going to get these questions again and again.
DR. APOSTOLAKIS: I must say, though, earlier on, somebody
said that we're not sure whether we're going to be -- this new 10 CFR
Part 50 will be better or worse than the existing one. And do you
really want to say that?
MR. CUNNINGHAM: There was a question to me, which maybe I
misunderstood the question, but it was will people be more or less safe
with the new Part 50.
DR. APOSTOLAKIS: And the answer was that they may be less?
MR. CUNNINGHAM: No. They're not going to be less safe, but
they may not be more safe either. They may be just as safe as they are
DR. APOSTOLAKIS: The word "worse" through me off. I hope
we're improving the process here.
MR. CUNNINGHAM: We're improving the processes. There are
things that we need to fix in Part 50.
DR. APOSTOLAKIS: We are improving the process.
MR. CUNNINGHAM: But we're not -- the goal of this is not
to, from the beginning, improve safety. That's not one of the goals.
DR. APOSTOLAKIS: I agree.
DR. WALLIS: No, but unless you have a measure of safety,
you don't know whether you've improved it or not. You cannot just make
DR. BONACA: This is almost starting from scratch. In
reality, what you are doing, you are, first of all, currently stating
that you provide reasonable assurance of adequate protection to the
public. That's something that I think you can -- or has been proposed
in the statement.
You are also then making changes to Part 50 within the
constraints of, in part, some concepts which are already in Reg Guide
1.14. So you have a controlled process to modify that.
In that context, do you need to remake that statement or can
you rephrase that statement in a way that it would more describe the
process you're going through, the one of modifying the current process
within certain controls so that you maintain reasonable assurance.
DR. APOSTOLAKIS: Well, one has to be a little realistic, I
think, and if you look, for example, at the in-service -- no -- the GQA
regulatory guide that was published last year, I think most people would
agree that that's an improvement over what we had before, but I don't
think anybody can really put any numbers to the result, because you are
just classifying components in a different way.
You are dealing with an issue that nobody can quantify the
impact of QA requirements on failure rates and so on. So there is some
element of truth to the argument that you can be improving the process
without necessarily quantifying the adequate protection levels and so
And I think with the ISI, although there you can have a
little more quantitative insights, again, there was universal agreement
that this is an improvement over what we had before.
So I guess that was Gary's point that we're risk-informing
the process without necessarily --
MR. HOLAHAN: I think it's fair to say what we're doing is
analogous to what was done in 1.174. When you use that reg guide, it
was just for license amendments as opposed to some other things we're
talking about here, but we're making the same conclusion.
The purpose of 1.174, which was quantitative and
qualitative, was to make judgments about providing reasonable assurance
of adequate protection of public health and safety.
That was done in some combined way of numerical analysis,
plus checking that defense-in-depth wasn't eroded and that there was
good engineering and safety margins, and it was that combination of
things that we said was providing this reasonable assurance.
DR. WALLIS: I think you have a problem, though, because
there is a significant body of opinion that believes that they're being
provided unreasonable assurance, we're making them unreasonably
regulated, way beyond the needs of public health and safety.
If you're ever going to back off in response to those
people, you're going to have to have some justification for it.
MR. HOLAHAN: I don't have any problem with that, but that
is the same issue that we dealt with in 1.174. I think you can deal
with that issue, how safe is safe enough and where are you comfortable
going, in a combination of numerical analysis and philosophical and
DR. BONACA: Just to complete my thought process from
before. Isn't it true you are trying to maintain reasonable assurance
of adequate protection?
MR. HOLAHAN: Yes.
DR. BONACA: You are not really starting from scratch. To
the degree to which you can utilize the concept, I think you will be
with all the difficulties you have in providing from scratch every time
reasonable assurance of adequate protection, which is almost like saying
you start from scratch. You don't. You don't.
MR. HOLAHAN: I agree with that, but once we decide to say
we're going to change, there will always be the question of have you
changed enough or have you changed too much. So you're going to have to
have some test to say where you stop is an appropriate point.
DR. KRESS: Let me wade in a little bit on the
quantification issue. You're not -- when you put together new rules and
regulations, you're not going to be able to quantify the effect on, say,
risk, which I won't equate to safety, until you can identify the plant
changes and the operational changes that result from that rule.
You're not going to be able to do it ahead of time. You're
not going to be able to say this is going to result in a safer plant by
this much or a less safe plant than this much until you can determine
what the outcome of the rule is in the plant design.
So basically you can't do what some of the people are asking
you to do right now, and that's to say whether or not it's going to be
safe or not.
I think you can do something like Mario Bonaca said and say,
well, by the process we've outlined, we know we're not going to deviate
so far from our regulations that we already have that we have defined as
adequate protection, that we're not going to deviate so far, that we're
going to do it in such a way that we have reasonable assurance we don't
throw you over some line of risk that's unacceptable.
So I think you do have to define some measure of what that
line is, you know, and throw it over, and that may be the way you
quantify it. And that may be the same numbers that are in 1.174, for
DR. APOSTOLAKIS: It seems to me you shouldn't put the first
sentence there, because everything else refers to the changes and the
first one presumably is being accomplished now with the current 10 CFR
MR. CUNNINGHAM: That's a good comment.
DR. APOSTOLAKIS: So you really don't need that there. It's
not that the change will lead you to a 10 CFR 50 that provides
reasonable assurance. That is already accomplished right now.
MR. HOLAHAN: We're providing an alternative.
DR. APOSTOLAKIS: An alternative, yes. But it seems to me
you can very easily start with include requirements and so on, and maybe
say something about the relative change regarding reasonable or adequate
protection or say nothing.
MR. HOLAHAN: Something makes me suspect that Dr. Wallis
will raise the same question, whether we --
DR. APOSTOLAKIS: Now that you put it there, yes.
DR. WALLIS: I think you do have, because when you get down
to maintenance of sufficient safety margins, you have to have some
criterion; now, what is sufficient?
MR. HOLAHAN: I don't have any problem with that. All I'm
saying is that when we get to that point, we ought not to expect that to
be a health effects calculation or a core damage frequency. It's going
to be a combination of concepts. There are valued safety concepts which
are not necessarily numerical risk calculations.
DR. APOSTOLAKIS: Let's go.
MR. CUNNINGHAM: The second point is that the new or the
revised Part 50 would include requirements on specific attributes of
design and operations commensurate with their safety significance. We
would want to write it so that the -- it reflects, if you will, the
plant-specific nature of some of the safety-significance of the design
and operational attributes.
We want a clear and coherent set of requirements, provides a
regulatory basis for all of our activities. It's performance-based to
the extent that we can do that and it's practical to implement.
The challenges, I think, or where we would most want to
change from the present Part 50 is in a couple of these really right
now. One is that today the requirements are not necessarily
commensurate with their safety significance. We know there are
requirements in Part 50, as somebody was kind of alluding to before,
overdo it, if you will; that there is conservatism there that we don't
need today, with our understanding today.
DR. POWERS: When you say that, I looked at several of these
regulations and one of them seems just way out of line with respect to
the others. How did you reach that conclusion?
MR. CUNNINGHAM: You see assumptions built into the
analysis, for example, that we could have a much more realistic estimate
of today, with a much smaller uncertainty, if you will. So the safety
margin that's built in there is large compared to what we think the
uncertainty is today.
DR. POWERS: I guess what I'm looking at is do you do you do
this out of a context of the plant during operations or the threats due
to fire or external events or shutdown operations?
MR. CUNNINGHAM: It has to reflect all of those things.
DR. POWERS: It has to, but does your assessment, the ones
out of line with the other, reflect all of those things?
MR. CUNNINGHAM: I haven't done the assessment yet. We're
going to come back to what we would -- how we would do the assessment,
but I think it does have to reflect all of those considerations.
DR. POWERS: Okay. Good.
DR. APOSTOLAKIS: Now, the three lines there that are
indented, consistency with defense-in-depth and so on, all this flows
from the safety goal policy statement, correct?
MR. HOLAHAN: No.
MR. CUNNINGHAM: No. No. They flow from 1.174.
DR. APOSTOLAKIS: Which flows from there.
MR. CUNNINGHAM: Which flows, in part, from the safety goal
MR. HOLAHAN: The 1.174, in fact, derives from the PRA
policy statement, which is a place where these different thoughts were
DR. APOSTOLAKIS: But even in the safety goal policy
statement, there are references to defense-in-depth and that the PRA
should supplement it and all that stuff, or is that policy statement?
MR. KING: It's in the safety goal policy. It seems to get
more refined as the PRA policy was developed, 1.174 was developed.
DR. APOSTOLAKIS: As you know, defense-in-depth if something
that is being discussed now in various quarters. Do you think that the
Commission might change the policy there at some point in the near
future, so that would change your approach?
MR. CUNNINGHAM: Change the policy with respect to
DR. APOSTOLAKIS: Regarding defense-in-depth and risk
assessment. See, today, the way I understand it, the risk assessment
really is not a principle. I mean, defense-in-depth is a principle. So
the PRA supports the defense-in-depth.
Now, what if they come out tomorrow and they say it's the
other way around, would you change anything here or do you think that
this is flexible enough to accommodate?
MR. CUNNINGHAM: I think this is probably flexible enough
and a lot of the information that would be used to make that decision to
change or not is probably going to come out of programs like this, as
well as work that's going on to relook at the safety goal policy
statement and that sort of thing.
These may be the stalking horses, if you will, to, if you
will, drive that type of change. I just don't anticipate the Commission
changing it independent of what we're doing in these types of projects.
DR. APOSTOLAKIS: Okay. Let's go on.
MR. CUNNINGHAM: There are two key points here. The present
Part 50, where we have troubles with it is, again, the requirements are
not today commensurate with their safety significance, maybe on either
side. Either there are too many requirements relative to the safety
they provide or too few.
The other part is the present Part 50 is not consistent.
It's not coherent, in part, due to the issue we talked about earlier on
reasonable assurance; in part, because of consistencies across the
They were developed over 30-some years, so there are
inconsistencies. That caused problems with the way Part 50 is today.
That's part of what we're trying to change.
MR. HOLAHAN: I'd just like to remind you that this slide is
a summary of the paper which the committee wrote a letter on supporting
six months ago.
MR. NEWBERRY: Eight or nine months ago or something.
MR. HOLAHAN: That's a historical fact.
DR. WALLIS: We're just checking that you understand what
you mean by those words.
MR. HOLAHAN: We mean the same thing you thought we meant
when you agreed with us.
MR. CUNNINGHAM: In the paper that the committee wrote a
letter on eight or nine months ago, there were three options for making
the changes to Part 50.
One was to continue with the existing rulemakings that were
underway at the time on the maintenance rule and 50.59, on reporting
requirements, that sort of thing, and, in a sense, stop there and make
no other changes.
The second option was to proceed to make changes to the
scope definitions used in certain parts of Part 50, under what we call
the special treatment, what we defined in the paper as special treatment
rules; for example, equipment qualification, QA requirements in the
The third option was to make changes to specific regulatory
requirements. This could be adding requirements, deleting requirements,
or modifying them to look at a more realistic analysis method, to change
the acceptance criteria, that sort of thing. So we went forward with
these three options.
We also went forward with a set of four policy issues.
First was should the conformance with this new Part 50 or modified Part
50 be mandatory on all licensees or should it be voluntary. Within
that, if it was voluntary, should licensees be allowed to select which
parts of it they want to conform with and which ones they don't.
The second policy issue was the potential for issuing
exemptions to pilot studies, to pilot plants, to help us understand
better how we should be risk-informing Part 50. Third was modification
of the scope of the maintenance rule. The maintenance rule, when
originally promulgated, had a fairly -- purposely had a fairly broad
scope identified in it.
Now, what we would be doing is reducing that scope. We
wanted that Commission decision to do that.
Fourth is a somewhat separate issue of clarification of
staff authorization. On it, we proposed to develop guidance of when the
staff can use risk information in the event that a licensee does not use
it and the definition of what's an acceptable risk under those
DR. APOSTOLAKIS: I'm a bit confused by this. Let's say we
have a situation where conformance with the modified Part 50 is
voluntary. Then the staff has the authority to use risk information in
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: What happened to the voluntary part? I
don't understand that.
MR. CUNNINGHAM: The last option, the last policy issue
deals with the staff use of it when the licensee chooses not to use it.
DR. APOSTOLAKIS: So then you are coming back and saying,
no, you will use it.
MR. CUNNINGHAM: We will use it, but what we're trying to do
in this fourth point is define how we can use it and what the acceptance
criteria are, if you will, for making the decision that even if the
licensee doesn't use it, here is our decision to use it.
DR. APOSTOLAKIS: So let's say you have a situation that
falls under one of the regulatory guides we already have and the
licensee chooses not to use 1.178. Then you are allowed to use risk
information and you will define a process for using that risk
information in decision-making, which will be different from what is in
that regulatory guide.
MR. HOLAHAN: It could be, but not necessarily. It could
DR. APOSTOLAKIS: But if it is the same, then essentially
you are telling the industry, sorry, but this is not voluntary anymore.
Aren't you saying that?
MR. HOLAHAN: Remember, what we're saying is there may be
special circumstances under which the licensee is proposing to do
something new and unusual that hasn't been looked at before and we'll
have to find a way to address the risk implications of those things.
That's when these issues tend to show up. It's not when
they want to use or stay with their current licensing basis. It's not
when they want to use standard technical specifications. It's not when
they want to use any existing reg guide.
DR. APOSTOLAKIS: I understand that, but isn't this a way of
essentially making conformance with a modified Part 50 mandatory? In
fact, it's worse, because now it's not mandatory. It's at the
discretion of the staff. Sometimes it's mandatory, sometimes it's not,
depending on what the staff decides.
DR. KRESS: I don't think so. If a licensee opts not to use
the modified Part 50, then the NRC says okay. I mean, there's no --
DR. APOSTOLAKIS: I know, but four will give them a license
DR. KRESS: They're not doing anything, though. When the
licensee says I'm going to stick with what I've got, the NRC is going to
MR. BARTON: I don't think so. I think the staff is still
going to come back and ask risk type questions.
DR. KRESS: If the plant is not going to conform to the new
DR. APOSTOLAKIS: No, no, no. I think there are two
different issues now. Tom is saying what if a licensee does not ask for
DR. KRESS: No. He's just going to stick with his --
DR. APOSTOLAKIS: Okay. He is allowed to do that. If he
asks for something --
DR. KRESS: If he asks for something, then --
DR. APOSTOLAKIS: Then four essentially allows the staff to
make conformance with a modified Part 50 mandatory, but at the
discretion of the staff.
MR. BARTON: That's what it sounds like.
MR. HOLAHAN: No. You're assuming that the same standard
would be used in item four as in other places, and that's not clear yet.
We haven't developed how that process would work and what that standard
would be. We're running then on the parallel track. We'll be back to
the committee in September, I think, with a recommendation on that.
DR. APOSTOLAKIS: The concern that I have is that, okay, you
say even though you don't use risk information, Mr. Licensee, I
naturally want to know what your change in the CDF is. Then they come
back and say, okay, this is a change. Then you say, well, naturally,
now I want to compare it with the guidelines from 1.174. You would be
crazy not to compare it.
So what's left? I mean, the qualitative part you do anyway.
So you're essentially applying 1.174, which was voluntary to begin with.
MR. CUNNINGHAM: But you're getting a little bit ahead, as
Gary was saying, that the criteria that would be used in this case is
something that was for a committee meeting here in September or
DR. APOSTOLAKIS: Another thing I'm saying is that I was
always uncomfortable with this voluntary versus mandatory conformance.
I just don't think it can be voluntary. Why don't we recognize that and
then instead of looking for ways for the staff to have authority to use
risk information, to find ways for the licensees to not use some of this
information, these regulatory guides in certain instances?
For example, when they don't ask for anything, you don't go
back and say, now, we want to make sure that you comply with the
risk-informed Part 50. But if they ask for some change, they request a
change, I just don't see how it can be voluntary.
I mean, even if you decide not to ask the question, we will
ask it. It's a CDF change.
MR. HOLAHAN: I don't think you'll ask it in every case.
DR. APOSTOLAKIS: No.
MR. HOLAHAN: Just as the staff is not going to ask it in
every case. We're going to look for cases in which this might be
DR. APOSTOLAKIS: But that's where the 1.174 comes into the
DR. POWERS: Let me understand, Gary. Suppose -- I can
imagine a couple of situations. Suppose a licensee asks for, say, a
power upgrade or something like that; that staff looked at it and said,
well, yes, true, it seems to comply with all the regulations, but I
know, from a risk assessment, that this is a very risky thing.
And they say, so, I'm going to, because I've gotten,
presumably gotten leave from the Commission to do so, I'm going to
reject this based on my understanding of the risk, and you write that
all down and you say I reject that based on my understanding of risk,
even though the licensee submitted nothing on risk.
That's one limit of where you're saying you might take
MR. HOLAHAN: Yes.
DR. POWERS: The other limit I can imagine is you've got
this same application before you, you have the same background
knowledge, and you write to the licensee and request for additional
information. You say, well, how did the CDF change? The licensee comes
back and says I haven't got a clue, because I've never done a risk
assessment on this, don't intend to, and I remind you that I comply with
all the other regulations, and so you should go ahead and approve it.
That's another limit. And don't you have to say, well, fair
enough, we know where he stands?
MR. HOLAHAN: yes.
DR. POWERS: And you make your decision based on whatever
you make your decision on having done that.
MR. HOLAHAN: Yes. That's exactly right.
DR. POWERS: And you have not compelled him to do and have
no license to compel him to do a risk assessment if he chooses not to.
MR. HOLAHAN: That's right. But it may affect the outcome
of the decision.
DR. POWERS: It may affect the outcome of the decision,
because you have leave to utilize whatever risk information you have.
MR. BARTON: I think the "may" is a "shall" or a "will"
DR. POWERS: I think they would be delinquent if they had
risk understanding of an issue and did not use that and I think they
would be delinquent if they had risk information that counter-indicated
a licensee's application and they didn't ask him about it.
If the licensee says I don't believe these PRAs, I think
they're a bunch of nonsense, I'm not going to use them for my license
application, and he might well respond to a staff indication that there
is a risk problem by indicating why he thinks it is not a problem and
still not take any response to doing a PRA, I mean, I don't see the
compelling nature of this.
The encouragement I definitely see in it, but that's part of
policy. The compelling nature I don't see, at least in those two
examples I picked.
MR. CUNNINGHAM: The point of the issue and going to the
Commission was we need to write down, so everybody understands, the
rules of that part of the game, if you will, and they are not written
DR. POWERS: Okay.
MR. CUNNINGHAM: Under what circumstances you can ask the
question and by what circumstances or what guidelines we can accept or
reject, based on our own risk analysis.
DR. POWERS: What you're asking is a constraint on your
behavior. That's a very unusual thing to do. You want to be careful
about that. They might give you an answer.
DR. APOSTOLAKIS: The first policy issue then refers to the
new Part 50, the risk-informed Part 50.
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: The licensee hasn't done anything.
MR. CUNNINGHAM: Yes.
MR. HOLAHAN: Yes.
DR. APOSTOLAKIS: Number four refers to requests for change.
MR. REED: Relaxations, yes.
DR. KRESS: For anything.
DR. APOSTOLAKIS: Is that what it is?
MR. HOLAHAN: Yes.
MR. CUNNINGHAM: Yes. Because if we see risk issues that
are not related to a request, it's always the backfit rule, if we have
new research, information or whatever, and you can always use that
DR. APOSTOLAKIS: It would really help a lot, I think, if
you made that clear. Clarify the clarification.
MR. CUNNINGHAM: Okay.
DR. APOSTOLAKIS: The fourth one really refers to requests.
DR. KRESS: Not necessarily. I think in our previous letter
on this issue, we said that we didn't think you had any choice but to
make it voluntary, because you'd have a horrendous problem with the
backfit rule otherwise.
MR. CUNNINGHAM: Yes, with respect to the first issue.
DR. KRESS: With respect to the first issue. The fourth
issue, I think, is just broad-based. When do you have authority to use
risk-informed processes across the board? I don't think it's limited to
DR. BONACA: What you're saying is that it cuts both ways.
MR. REED: Yes, it cuts both ways.
DR. BONACA: The information is being used and it cuts both
ways and it has to be that way.
MR. REED: And it's also applicable to the current Part 60.
DR. APOSTOLAKIS: But the truth of the matter is that if
somebody says I don't want to have anything to do with the new process,
please leave me alone, I have met all the requirements, I want to
continue under the old system, the truth is that that licensee cannot do
MR. HOLAHAN: Well, if that licensee says I want to continue
in the old system, but I want to change my licensing basis, it opens up
the legitimate questions about what are the risk implications of these
DR. APOSTOLAKIS: So it is really related to requests.
MR. HOLAHAN: Related to requests.
DR. KRESS: I guess it is, yes.
DR. APOSTOLAKIS: Then I understand.
DR. BONACA: What if it changes the licensing basis, not
based on PRA insights, but purely some tech spec change or whatever?
MR. HOLAHAN: Yes.
DR. BONACA: Would you still --
MR. HOLAHAN: Well, what we're talking about is probably in
order to do item four, the first thing you need to do is to develop some
screening process to say, you know, what kind of power up rates are very
unlikely to have any risk implications, and then you do those through
the normal process.
But if you get ten or 20 percent power up rates, you may be
introducing risk considerations even though you meet all the other
So what we're looking for is probably, first, a screening
criteria. Maybe it's only one out of 100 cases in which this
application by a licensee needs some special risk assessment. All we're
saying is that we want to clarify the rules of the game so that we know
and the licensees understand, if they send us something that has risk
implications, how are we going to deal with it.
DR. BONACA: I have just one more question. Assuming that
you go voluntary versus mandatory conformance, could you have a
situation where somebody says I'm not going to go to modify Part 50;
however, I have a risk-informed application for a specific element, and
so I'm going to submit that one.
So I'm still living under the old Part 50, but I want to
have this. Can he do that?
MR. HOLAHAN: Yes.
DR. BONACA: But if he can do that, then you're going to
MR. HOLAHAN: But I might ask you an item four question if
it's an unusual or --
DR. BONACA: I understand that. But I'm saying that you're
going to have infinite variations of mixes of all Part 50 -- you have a
series of cocktails that's pretty amazing, and I'm trying to understand
what the future is going to be.
I mean, how are you going to track that?
MR. CUNNINGHAM: It's going to be complicated.
MR. HOLAHAN: It's going to be complex.
DR. BONACA: Complex, to say the least. And what about the
burden for the staff, for example, to try to control an environment
where you have this mix, because you have that kind of situation. I'm
sure a lot of people will say I want to stay with the old Part 50, but I
like the removal of hydrogen control.
DR. APOSTOLAKIS: I guess we can raise those questions when
item four will be discussed in the future. You don't have a proposal
MR. CUNNINGHAM: No, we don't have a proposal on it.
DR. APOSTOLAKIS: So we will come back to this.
DR. BONACA: Although, if you want to talk about it that
way, it really blows my mind now, and then I always thought it would be
voluntary, but in this kind of backdrop, I begin to have a lot of
questions, in my mind.
DR. APOSTOLAKIS: At the higher intellectual level, it's
really crazy to say I choose to use this piece of information and I
choose to ignore it.
DR. BONACA: Absolutely.
DR. APOSTOLAKIS: The information is there. I mean, it's
there. I mean, it affects your state of knowledge and when you make
your decision, right? How can you ignore something when it's there?
DR. WALLIS: You can do that in the regulatory world.
DR. KRESS: Well, the information may not be necessarily
DR. APOSTOLAKIS: Oh, it is there. It's a matter of
DR. KRESS: Well, if you think the IPE is the information.
DR. APOSTOLAKIS: If the staff can calculate it, then the
licensee can calculate it. So, in my mind, it's there.
DR. KRESS: My point is the staff probably cannot calculate
it. They don't have the plant specific PRA and if it's some change, the
licensee may not have calculated it. So the information may not be
DR. APOSTOLAKIS: I think it is there. They feel it's
important enough to quantify it. It's there in the sense that somebody
can sit down and maybe in six months can produce it. But it's there. I
mean, we're not asking them to quantify something that we don't know how
to do. You can always calculate, in the serious cases, the impact from
CDF. It may take you a while, but you can do it.
DR. KRESS: Who can?
DR. APOSTOLAKIS: The licensee.
DR. KRESS: The licensee may not want -- may say he's not
going to, though. That's my point. It may not be there. Then they
can't do it because they don't have the plant specific.
DR. APOSTOLAKIS: Physically it may not be there, but what
I'm saying is that conceptually, it is there, and we choose not to use
it, which is really a crazy approach. Okay.
MR. CUNNINGHAM: At any rate, the next two slides cover what
the Commission told us in response to our proposals. Basically, in a
nutshell, they said go ahead with options one, two and three, in
DR. APOSTOLAKIS: Right.
MR. CUNNINGHAM: Option one, go ahead and continue the
current rulemakings. Option two, develop a rulemaking plan for
addressing the special treatment rules and changing the scope and
definitions; you can use pilot plants, with exemptions, as part of that;
go ahead and change the scope -- as part of this, change the scope of
50.65 to conform with the rest of the rules. And to proceed with option
three, which is to study the potential for other regulatory changes to
DR. APOSTOLAKIS: That confuses me, though. These are not
options really, are they? If you can approve all of them, they're not
MR. CUNNINGHAM: They're options.
DR. APOSTOLAKIS: I got confused. I thought I didn't
understand the vote.
MR. CUNNINGHAM: No.
DR. APOSTOLAKIS: We approve three, but we also approve one
MR. CUNNINGHAM: Yes.
DR. WALLIS: I was confused, too.
DR. APOSTOLAKIS: Yes. I mean, the word "option." Okay.
MR. CUNNINGHAM: Options was, in retrospect, not a good
choice of words.
In terms of the policy issues, the Commission said we should
assume that this will be a voluntary modification to Part 50. We cannot
mandate that all licensees change to the new Part 50. They talked about
the issue of selective implementation, where, to make it even more
complicated, if somebody chooses to go to the new Part 50, can they
choose to institute part of the new Part 50 and not all of the new Part
We talked about that with the Commission and the Commission
said it's premature to make that decision, basically; so come back when
you've done some more work and talk to us about that, if you will.
DR. KRESS: They weren't all agreed on item one, either.
Did they? Item one was also not unanimous.
MR. CUNNINGHAM: I don't remember.
DR. KRESS: Voluntary versus mandatory.
MR. CUNNINGHAM: I don't remember. But the Commission
decision was voluntary.
DR. KRESS: The Commission decision was to go to voluntary.
MR. CUNNINGHAM: Yes. The Commission approved the use of
industry pilot studies, with exemptions. They approved a modification
of the scope of the --
DR. WALLIS: The pilot studies, have they said anything
about how you measure what you learn from these studies?
MR. CUNNINGHAM: I'm sorry. I only caught part of what you
DR. WALLIS: Pilot study is an experiment and you have to be
measuring something with this experiment. You don't just sort of
describe it like a journalist when you've done it. You actually look
for some meaningful results. This gets back to the quantification issue
and what is it you're going to measure.
DR. POWERS: To quote the EDO, the pilot studies aren't to
test concepts, they're only to find out if things work.
DR. WALLIS: How do you know if it works?
DR. POWERS: You have to ask the EDO that question, and I
think you will get a chance to.
DR. BONACA: But the evaluations always conclude whether you have a risk
neutral or risk reduction, is it?
MR. CUNNINGHAM: Yes. But what you don't get absent pilots
is some indication of whether, as Dr. Powers was saying, that you get
some real benefit from this modification; is it worth changing a
particular rule if it doesn't really improve the clarity or the process
side from our side and it doesn't save the licensee any money.
MR. HOLAHAN: But you also get some information on the
practicality of what's been proposed and on at least some potential
unintended consequences that may show up. So you understand better the
long-term implications of putting some of the things you're testing into
I don't think it's really going to be a test of are you --
how much is safety changed. It's probably much more a test of the
practicality of implementing the concepts that are being put out.
DR. POWERS: I think that the concern that has arisen before
in the committee is certainly the quantification Professor Wallis brings
up, but also the interpretation of the results. When you do an
experiment, you clearly perturb the system and these pilots tend to get
the focus of the best and the brightest, and a group of motivated people
anxious to make things work well.
And you hear about an overly optimistic result coming out of
them that does not translate well. You worry about the best and the
brightest applying something that the rest of us are not skilled at
And I think what we telegraph to you and we have, in a
previous letter telegraphed to you, is especially in these areas where
you're talking about fairly radical things, that, yes, layout some
specific objectives, quantify it, do the best you can, so that you can
answer the questions that Professor Wallis brings up -- have you really
done what you think you've done -- and talk to us some about the
uncertainties and perhaps over-optimistic interpretations you might have
on these pilots.
I think there was a time when pilots seemed to me to be
relatively unusual events and now there doesn't seem to be a single
thing that comes up that doesn't have the word pilots attached to it.
I think we have to be a little more circumspect in how we
apply them and not just say, well, we've had a pilot and assume that
that somehow anoints this with the aura of respectability, when, in
fact, you may not have shown anything at all in the pilot.
Is that a fair summation of all the concerns about pilots,
DR. WALLIS: Well, I feel you're rather like the professor
trying to explain the words in a different way to see if the student
DR. APOSTOLAKIS: Maybe we need to --
DR. POWERS: One of the reasons I have no interest in
becoming a professor, I guess.
DR. APOSTOLAKIS: Maybe we need a little project that you
guys should undertake one of these days on the design of pilot projects
and have some broad guidelines and maybe a framework, which does not
have to be quantitative in every single respect.
MR. HOLAHAN: A pilots pilot.
DR. APOSTOLAKIS: Then you have a pilot to test the pilot
designs, but that certainly would help a lot, because I agree with Dana
that this is now something that we are doing routinely and we don't seem
to think as if there is no evidence, I'm sure you guys are thinking
about it, that there is no evidence that there is some systematic
thinking about the pilots beforehand.
I'm sure there is -- as I said, you think about it, I mean,
definitely, and you worry about certain things, but the appearance is
that this is done in an ad hoc way and it wouldn't be much of a deal to
really put it down on paper and refine it to the point that it will
satisfy the people who are more skeptical.
MR. HOLAHAN: I think that's a very legitimate issue and
maybe we can do something on that. But part of, a very substantial part
of the pilot activity is, in fact, the people who are volunteering and
as a practical matter, most of the pilot activities, at least that we've
been involved with and discussed with this committee, are things that
licensees volunteer to do and we have, sort of in a reactive sense,
accepted them as being reasonable ways of testing pieces of the
questions we are interested in.
And clearly they are not always structured to do exactly
what we might want to test. They're not always complete. They don't
test every aspect, because almost always, by their nature, these pilot
activities are voluntary. You can't always control that well the people
who are volunteering. So you accept what they offer and you get what
best information you can out of it.
DR. APOSTOLAKIS: Yes. But surely, I mean, if the licensee
sees a document that's well reasoned that says this is what we would
like to learn from this pilot, I mean, the technical part, nobody can
object, unless they feel that some parts of it may be will impose
additional excessive burden.
But this is something you can negotiate with them.
DR. BONACA: And it seems to me that 1.174 gives you a lot
of control on the approval and evaluation, doesn't it?
DR. APOSTOLAKIS: Not for the pilots, though.
DR. KRESS: Not the pilots.
DR. APOSTOLAKIS: Not the pilots. The pilots are completely
MR. NEWBERRY: Well, I think the principals and the thought
process could be applied to many of these issues.
MR. BONACA: The way I see some of this, the self-evaluation
provided the support for those that seemed to spring out directly from
DR. APOSTOLAKIS: You are still on page eight, Mark.
MR. CUNNINGHAM: Yes, sir.
DR. APOSTOLAKIS: You are really going very slowly.
MR. CUNNINGHAM: I apologize to the committee for that.
DR. KRESS: It's because he talks so slow.
MR. CUNNINGHAM: If I could finish the last point on page
eight, which is the Commission approved us to go ahead and develop the
guidance saying go back in September.
DR. APOSTOLAKIS: This is new, right? We haven't seen this
MR. CUNNINGHAM: Yes, that's correct.
DR. APOSTOLAKIS: That's correct.
MR. CUNNINGHAM: Correct, you have not seen this before.
This lays out the organization given that the Commission had approved
option two and option three to run, in effect, parallel. What you've
got is, in effect, two parallel organizations working the problem.
On the left-hand side, you have the option two part of it.
NRR has the lead for that. On the right, you have the option three
part, which Research has the lead on. Over top of each of those, you
have two levels of management review here on the picture.
The risk-informed licensing panel is a collection of
division directors who are going to try to work a lot of the issues that
come up and coordinate what happens between option two and option three.
DR. WALLIS: I don't understand why one option is assigned
to NRR and one to Research. I don't see what this has to do with
Research. It seems to me you're trying to get regulation. I'm not sure
what this has to do with Research. It seems to be all NRR's job.
MR. CUNNINGHAM: A lot of the work that will end up being
done as part of -- it's anticipated that a lot of the work that will
done in option three is the revisiting on the new development of a
technical basis for specific rule changes. That development of a
technical basis is the responsibility or Research.
DR. WALLIS: I could see Research providing support.
MR. KING: When we get into the details, it will become a
little more obvious.
DR. WALLIS: There is a Research task?
MR. KING: Research has the lead for this.
MR. HOLAHAN: Did you mean Research with the small "r" or
Research with a capital "R?"
DR. WALLIS: I'm trying to determine the difference between
the Research branch of the agency and the Regulatory branch.
MR. KING: Remember, option three is a study. Research has
the lead to do the study. What comes out of that study will likely be
recommendations for rulemaking. At that point, there will be a handoff
DR. WALLIS: So it's a more preliminary state.
MR. KING: Right.
DR. WALLIS: You're not going to actually make changes.
MR. KING: No. Research won't make the changes. There will
be a handoff to NRR to do that. You'll see that later in the viewgraph.
MR. HOLAHAN: It's analogous to the generic issue program in
which Research studies an issue and makes a recommendation about whether
there should be a new requirement. That's passed off to NRR to do the
DR. WALLIS: But I think there is a difference. Sort of the
administrative part of the agency has got a job to do, they do it,
they're asked to risk-inform something and they do it. Research is
going to be more at the level, I hope, of asking questions about do we
do this at all, what do we really mean by it, how do we measure things,
all that kind of stuff.
MR. HOLAHAN: Yes, that's right.
MR. CUNNINGHAM: It's making recommendations and that sort
of thing. The implementation of the rule changes will be NRR's
Over top of the two groups, again, is the risk-informed
licensing panel, first. Above that, for more of a policy type of
overview, is the PRA steering committee, that's basically composed of
DR. WALLIS: What is this RIP-50? Rest in peace 50?
MR. HOLAHAN: I haven't noticed any rest and I don't expect
MR. CUNNINGHAM: Implicit on this, by the way, is the fact
that policy issues will come up inevitably. We're sure policy issues
will come up. There's a set of those policy issues that will have to go
to the Commission. So implicit on top of all of this is the Commission
and the EDO for their decision-making.
DR. APOSTOLAKIS: So who is the lead? Is there a single
person responsible for option two?
MR. NEWBERRY: Let me answer that. DRIP, right there,
regulatory improvement program, we're the outfit that has led and had
the lead for 50.59, FSAR. This is being handled the same way. We are
the lead. The project management function in NRR to coordinate with all
the technical divisions and the different people here to complete this
If the Director of NRR has a question on this activity, he's
going to come to Dave Matthews or Scott Newberry in DRIP and we'll have
to get the answer for him.
DR. APOSTOLAKIS: And how about Research?
MR. CUNNINGHAM: The parallel is --
MR. KING: Is me. If Thadani has a question, he comes to
DR. APOSTOLAKIS: Okay.
MR. NEWBERRY: Let me -- constructing this diagram was no
small task, since we got the SRM, and various options were considered.
An option was proposed somewhere for a dedicated task force, things like
that, to create a separate organization, and we thought about that and
decided -- we recommended to the steering committee that we not do that.
We felt that we should retain the responsibilities in the
organization and truly struggled with the risk-informing of the
regulations that are in the line. It's just in the longer term, you
would not get the payoff and the buy-in and the understanding of
changing the fundamental regulations and the regulations that are in
So that's why we've constructed an organization like this.
There will be engineering staff, there will be inspection staff, there
will be project management staff all involved, depending on what
regulation is being changed and their staff being -- I'll use the word
dedicated, a team effort here. Tim is one of the key project
managers on that team.
DR. WALLIS: It looks pretty -- I can't evaluate it at all
because I don't see a sort of breakdown of the tasks and I don't see why
you need to do it this way. So I can just say it looks pretty.
Presumably someone looked at the job and figured out what needs to be
MR. NEWBERRY: We'll get into some of the tasks in the next
DR. WALLIS: In making up this diagram.
DR. APOSTOLAKIS: Move on.
MR. CUNNINGHAM: I'll turn it over to Tim now to talk about
the details on option two.
MR. REED: Okay. This is our plans right now for option
two. What you can see in the overview, we have planned four tasks that
we want to build into rulemaking plan, that we're on the hook to give to
the Commission at the end of October this year.
I'd preface this first by saying these are just our thoughts
of this core team that you just saw in that diagram, the slide before,
and I guarantee you that the next time you see it, it will be different,
because the more you think about the stuff, the more you find it's a lot
more complex and things change quite a bit.
In the overview -- and I have a slide on each of these tasks
that will follow. The way we see it breaking out right now is the first
task, of course, will determine which rules in Part 50 are these
so-called special treatment rules that we're going to revise the scope
for, risk inform the scope of these rules.
So first of all, we've got to determine what's in option
The second task, we broke out one rule which is difficult, a
burdensome rule, and it's complex because the scope is intertwined with
the body of the regulation itself, and we have to look at this one by
itself, and that's 10 CFR 50.59. So we broke that out as a separate
DR. APOSTOLAKIS: Yes. We haven't looked at that in a long
MR. REED: I understand quite a few people have lined up and
talked to the committee on that. It's one that's near and dear to my
heart, from my past life. But anyway.
Task three then would be really -- it should just say that's
the rulemaking approach. What that is is how are we going to go about
actually doing this in Part 50. As you'll see, we're going to use some
new terminology and our plan is to try to phase it in rule by rule.
DR. WALLIS: Excuse me. You're making decisions here.
You're determining which rule is to be considered, the best approach.
Presumably, you're going to have some sort of a measure. You're going
to look at the rules and have some measure of whether they're worth
MR. REED: Yes. Actually, when we get to this slide --
DR. WALLIS: And what's the payoff for doing it. The best
approach has got to be measured on some scale, so we're getting back to
MR. REED: It will certainly be -- no. I was going to say
qualitative, but we'll do what we can here.
Let me just give this overview here, and then you can take
shots at me on each of the tasks I have a slide on.
DR. WALLIS: I'm just saying it would be nice to see, when
you do this, some very clear criteria about how we determine the best.
MR. REED: Actually, we have to develop criteria to
determine what's in scope and what's out of scope. We've recognized
that, first of all, because we have to defend, if something is not in
scope, why is it not in scope.
DR. APOSTOLAKIS: I remember that we had the presentation or
an earlier effort to risk-inform 50.59, headed by Mary Drouin, and there
was a list of criteria there, as I recall; not quantitative, but there
was a list.
So presumably you guys will do the same thing or better,
MR. REED: Absolutely.
DR. APOSTOLAKIS: Do better.
MR. REED: Always try to do better.
DR. APOSTOLAKIS: Okay.
MR. REED: Then the last task there is we attempt to try to
identify the requirements that would govern the method for identifying
this risk-informed scope of structures, systems and components, and I'll
talk a little bit more about that later.
DR. APOSTOLAKIS: I'm a bit confused here, Tim. It seems to
me that task four really should be task one or two. Is it not?
MR. REED: Actually, wait till I get there, but basically
what were thinking of, jumping ahead a little bit, is maybe having, for
example, a new definition in Part 50.2, called something, and then that
would be supported by, say, an appendix to Part 50 that would go about
describing a box or a set of requirements that would tell you how to
actually go about risk-informing or looking at the systems, structures
and components and deciding what is the real scope.
It would talk about things like PRA quality, whatever that
is, how much defense-in-depth you have to have.
DR. APOSTOLAKIS: Are these tasks to be pursued in parallel?
MR. REED: Yes. This is all going on at the same time.
Basically, we have developed this enough to make a rule-making plan,
that's what we're thinking now, and, of course, that's the immediate
product rulemaking plan, but we're thinking about it for the whole
effort, too, of course.
Anyway, let me go to a slide on each task now. Task one,
then, determine which of these so-called special treatment rules will be
in the option two effort. We think we need -- I think it's obvious we
need to develop some sort of criteria that defines what's in and what's
out. We do not have a criteria at this time.
What we have here, and you will see that I'm kind of jumping
ahead a little bit in terms of you're going to get some of the answers
to task three here real quick. But what we've done is we've kind of --
if it looks like it should be in, it should be in, kind of thing. This
is really the thoughts of the team at this point over the kinds of rules
that we think are in scope for option two right now, and they're on this
slide and the next slide, and I've grouped them into two bunches.
DR. APOSTOLAKIS: Now you have new terminology. You haven't
talked about phases yet.
MR. REED: Right. That's actually -- I'm combining two
things on the one slide here, unfortunately. I should probably not have
DR. APOSTOLAKIS: Can you tell us real quickly what is a
special treatment rule? What is that?
MR. REED: The special treatment rule, if you read from
SECY-98-300, and most people in this room are probably more familiar
with that SECY than I am, but it talked about treatment of quality, of
reference quality, and then it went to a parenthetical that said that
special treatment refers to things like quality assurance, 50.59, tech
specs, ASME code, and it listed those regulations as regulations that
apply special treatment to SSCs that are --
DR. APOSTOLAKIS: So they're safety-related.
MR. REED: Yes. Safety-related equipment, if you will.
DR. POWERS: All the things that don't fit well with PRA.
MR. REED: It's a lot more stuff than that.
DR. APOSTOLAKIS: What?
DR. POWERS: All the things that don't fit well with PRA.
MR. REED: Basically, things that get on the Q list at a
plant get a lot of care and feeding, some which may not be really
required or should not be required, I should say. It is required today.
So these are the kinds of things we see. For example, one
that's not there, obviously, the first one, new definition, we think
we're, jumping ahead again, we're already assuming we're already
assuming, we've already concluded that we think we need new terminology.
We don't even want to go basically try to redefine something that's been
used for the last 30 years, it's got a lot of baggage with it, that
creates a lot of confusion. We think it's a lot better approach, and I
believe industry does, too, to use actually new terminology and try to
stay away from creating that kind of confusion.
So we'll probably have a new definition. It will probably
be in 50.2, I believe the section is. We also see these regulations
right now listed there as special treatment regulations that we would
try to do perhaps in a first phase and, again, task three will talk
about why we believe it's -- we think it should be phased in, instead of
doing all the regulations at once.
DR. WALLIS: Excuse me. Are you going to do all these as
MR. REED: That's the concept at this point, yes.
DR. WALLIS: Have you done one?
MR. REED: I mean, this could change very quickly.
DR. WALLIS: I would be tempted to say do one and do it well
and show it worked.
MR. REED: Yes. Like the maintenance rule or something like
that. Yes. These are very preliminary thoughts.
DR. APOSTOLAKIS: These phases you're talking about are
related to the NEI phases. I saw somewhere four phases. Or are these
MR. REED: I don't believe -- yes, these are different
phases. These are phases that -- I didn't make any effort to try to
correlate them with NEI.
DR. APOSTOLAKIS: I'm really perplexed why you didn't ask
yourselves the question that Dr. Wallis just asked.
MR. REED: Actually, we did.
DR. APOSTOLAKIS: It looks like a big job here.
MR. REED: It's a huge job and we will do things in the
first phase, the rules that are simple to do, things that perhaps the
pilots are doing, we've got some experience from doing, things that fit
together well, things that give a lot of bang for the buck, perhaps, and
things that are more complex and more difficult and less experience we
would save to phase two, if you will, or whatever phase.
After we get some experience, we get better at doing this.
And we may have more than two phases and we may do one regulation at
first. Again, this is just our thoughts.
DR. APOSTOLAKIS: Are these five items you have there
candidates for evaluation or all of them you -- the intent is that you
would do all of them?
MR. REED: I think candidates is a better word, actually.
We're not that sure about --
MR. HOLAHAN: There is a difficulty -- it would be easier to
pick one and do it, but there is a difficulty in doing that. What we
would like is to have a definition and approach that ultimately apply to
all of these and maybe some others. And when you do them one at a time,
what you might find out is that the third one doesn't fit very well with
the definition and the approach that you established the first time.
So in order to make these things work well together, and
these are all rules that have to work well together in an integrated
sense for plant operation, I think you have to -- when you first begin
to look at definitions and using risk information for equipment
qualification, you have to worry about the implications of that on other
So we might like to do them one at a time, but there are
potential problems with doing that.
MR. NEWBERRY: Let me just -- maybe Gary said this, but he
can correct me. But I think the South Texas folks will be talking to
you eventually and their experience with the graded QA program, where
we, through considerable effort on their part and the staff's, we
approved their proposal. Yet they did not get the benefit from that
approval that they had hoped nor we had hoped.
And I think that lesson is a large driver here in that we
need to look in an integrated way, which is what Gary was saying,
because if you just modify one regulation which affects that specific
component, it could be a series of other multiple regulations that cause
the utility to invest in carrying of that component.
So we're going to have to look at the whole suite to see
what makes sense here. My feeling right now, it's going to be a rather
broad set of regulations that we're going to have to tackle here or it's
not going to be worth doing.
DR. WALLIS: While someone should look at the broad picture,
I agree, it seems that each one of these is a fairly big job. So that
while some group is looking at it broadly, drafting out everything,
saying how do you make consistency is going to be a major effort on
getting something finished.
DR. APOSTOLAKIS: Wouldn't you take -- speaking of South
Texas, the GQA, the QA and start with that and see what other
regulations are affecting it and then go there and go there and
eventually develop a network of rules that are in harmony?
MR. REED: Actually, that's a lot of the thought here. When
I put those tasks out there, I could have put pilot plant beside every
one of them, because we're going to be looking at what South Texas is
doing and also what the industry is doing, of course, and trying to
understand the interrelations between these different regulations.
South Texas, as just was mentioned, ran into problems in
graded quality assurance, the ASME code and 50.55(a) and other
regulations tied them up and basically stopped them from fully
implementing it. That's the kind of thing we need to understand, so we
can get those regulations out of the way and that our people aren't
tripping over them trying to do it.
That's what I mean by kind of grouping them together, those
that fit together, so you can implement a coherent piece that can
actually do it. That's the idea conceptually, and I agree, each
regulation is very difficult and very complex. We've already reached
MR. HOLAHAN: You know, I keep coming back to this. One of
the things we've discovered is, the problem is each of these regulations
was developed at a different time, by people trying to do the right
thing, and developing a different scope, different approaches, which, in
some ways, conflicted with each other, and I think in order to avoid
that, we need to keep our thinking on all of these regulations as
integrated as possible.
So from my point of view, what I would like to see is to
attempt to do as many as possible which are closely related. And when
we find that there are some regulations which are perhaps not so closely
related, perhaps most of Part 100 doesn't need to be risk-informed at
the same time, maybe just a little piece of it.
We should avoid dealing with related issues at separate
times and what we're trying to do is lay out these regulations that seem
to be closely related and do them together.
DR. APOSTOLAKIS: I think that perhaps what you need is
instead of showing these little arrows there, to show some logical
MR. HOLAHAN: Relationships.
DR. APOSTOLAKIS: Relationships. That might be a more
convincing argument. But you are looking really at defining something
that will replace safety-related classification, right? Is that what
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: You may have risk-significant,
risk-insignificant, and something in between, or four classes.
Now, is it reasonable -- maybe it's not such a big task then
if you only do that.
MR. HOLAHAN: If we find that the same concept applies to
all these regulations, then it's not so difficult. We want to test that
DR. APOSTOLAKIS: I understand. So why wouldn't you use
them Fussel, Vesley and RAW to do it for all of these? What is it
that's stopping that? Is that what you're going to test?
If I use these and I did it, would I have any problems with
50.49 or 55(a)? That's really the question you are asking. You already
know of one way of doing this.
MR. HOLAHAN: Yes.
DR. APOSTOLAKIS: You are just not sure whether that way is
applicable to all of these and that it will not create new problems.
MR. HOLAHAN: I don't want to endorse Fussel, Vesley and RAW
as a way of doing that.
DR. APOSTOLAKIS: I'm not endorsing it either, but that's a
way that has been used.
MR. HOLAHAN: Well, that, no. What has been used is that in
combination with expert panels and all that.
DR. APOSTOLAKIS: It's never risk-based.
MR. HOLAHAN: Right.
DR. APOSTOLAKIS: I know, I know. But, I mean, that
MR. HOLAHAN: Yes.
DR. APOSTOLAKIS: So you have a starting point.
MR. HOLAHAN: Yes. Certainly. Yes.
DR. APOSTOLAKIS: So it's not as big a job as I thought it
MR. REED: Actually, I think technically, it's a lot more
straightforward than doing the legal changing to Part 50 and all of the
DR. APOSTOLAKIS: I forgot that this is only task one.
You're not risk-informing each one of these. You're just looking at the
MR. REED: The scope. That's the idea. The second slide,
the next slide then is some more regulations that, for the time being,
we see as more difficult and see these coming in later. Again, this
could move around. As we get smarter and figure out which regulations
are really intertied with each other.
And, of course, pointing out a separate task is 50.59, and I
have a slide on that coming up.
And the reason we pointed this one out is this one was
pointed out in the SECY. This is a very difficult one to risk-inform,
as I'm sure this committee is well aware, and the reason being is the
scope of 50.59, what's described in the FSAR is throughout 50.59, as
well as words like previously analyzed accidents.
So once you try to start changing the scope of systems,
structures and components to which 50.59 applies, it starts unwinding
and pretty soon you end up changing the whole regulation.
So if we can do it in a clever manner, some clever simple
scope change and get some benefit, we'd like to do that, of course. So
that's a potential option.
But if it starts unwinding and we have to basically do the
whole rule, then you've got the second option there, I see simple
options to deal with this regulation.
DR. WALLIS: It would be the same intent, the same purpose.
MR. REED: Yes, but basically you find you have to do the
DR. WALLIS: But the fact that you may have to redo the
whole thing --
MR. REED: That's right.
DR. WALLIS: -- may be a good thing.
MR. REED: Perhaps.
DR. WALLIS: It would be consistent.
MR. REED: It could be a good thing.
DR. WALLIS: It wouldn't give you terms which were undefined
and could be argued about forever and that sort of thing.
MR. REED: Not to get into 50.59 too much, but I --
DR. WALLIS: I think it would be beautiful to streamline
MR. REED: 50.59 is so intertwined with the actual Chapter
15 accident analysis and the words mean things for that and only that,
that I don't know how you do this, to be honest with you, because risk,
almost by definition, goes beyond Chapter 15, multiple failures and the
So we'll have to look at it and see what we can do.
But the last option then is another more expedient approach,
if you will, and this I think is line with what South Texas is actually
asking for right now, and we'll see if that turns out to be the case.
But if you, for example, look in terms of the final product,
you have a Q list at a plant. Basically, what they're doing is going to
take some things off the Q list and make it a smarter Q list, is what I
like to view it as. You'll know exactly why something is on the Q list
and if it doesn't need to be on there, it will come off.
To do that, though, a lot of the stuff is described in the
FSAR and you're going to need pieces of paper for all this stuff, 50.59,
and you're going to have to kill a lot of trees, it could be, because
the Q lists are gigantic.
So you may want a one-time exemption to 50.59 to allow these
things to be reclassified and taken off the Q list. So it's that kind
of thing. It's more of an exemption type approach, not what we're
thinking in terms of actually changing the scope.
So I put it on there as an option. I think it could be very
well what happens in the short term.
MR. SIEBER: It's my understanding that the PRA standard
that ASME is working on is not approved yet.
MR. CUNNINGHAM: That's correct.
MR. SIEBER: And so the quality of licensee PRAs probably
goes from 100 down to some lower level.
MR. MARKLEY: Sixty-five.
MR. SIEBER: And you're going to rely on licensing PRAs for
all this reclassification about the standard. Wouldn't developing and
approving the standard be the very first step in the process?
MR. CUNNINGHAM: When we get to the issues, that's PRA
quality is one of the issues we have to face early on in this whole
MR. SIEBER: It would seem to me that that should be the
MR. REED: Actually, we have it as one of our tasks here
you'll see coming up. We mention it. We know it's a big -- it's an
DR. APOSTOLAKIS: So this is not risk-informing 50.59.
MR. REED: This is trying to risk-inform the scope of 50.59.
DR. APOSTOLAKIS: So when are we going to risk-inform 50.59?
Is that another project?
MR. HOLAHAN: Remember the way the options are set up.
Option two is supposed to become more -- make the regulations more
risk-informed by changing scope, not technical content, yet, and option
three should identify those regulations where we would change technical
I think what you heard Tim say is that we're finding it
difficult to talk about 50.59 in terms of just changing its scope. In
fact, in order to risk-inform 50.59, you may be forced to change not
just the scope of what it applies to, but to how it does its tasks.
We may find that we're forced to pull 50.59 as maybe the
earliest example of risk-informing the content of the requirement, as
well as the scope.
DR. APOSTOLAKIS: So that would be from option three, then.
MR. HOLAHAN: In effect, it would be instantaneous transfer
from three to two. We start out seeing how much can be done in the
scope context and whether that makes sense or not.
MR. REED: Task three, this slide could be much simpler just
saying it's the rulemaking approaches and really not determine, I guess,
more develop the approach for revising the terminology of Part 50 and
how to basically insert that into Part 50 and get it done; how do you
actually do this.
We've already evolved a little bit, I think, in our thinking
and have already determined that we need terminology, whatever that is.
We throw some words out, like essential equipment or safety-significant
equipment or SSCs or risk-significant components, or whatever words you
want to come up with.
We've got to be careful, of course, because there's a lot of
words out there that mean stuff and meant stuff for the last 30 years.
So this is a lot more difficult than it appears, believe it or not, just
to come up with the terminology.
DR. APOSTOLAKIS: That's why we're just aliatory, by the
way, in a different context.
MR. HOLAHAN: Yes, that was very helpful.
DR. APOSTOLAKIS: That was the reason, but everything else
was used in different courses, random, uncertain, I mean, you find those
words everywhere. So maybe you need a good Greek and Latin dictionary.
MR. REED: Then, also, as you've already seen earlier on,
we're going to evaluate the option of doing everything at once, more
phasing in, and you can obviously tell that we're thinking in terms of
phasing in by regulations and perhaps groups of regulations that make
sense to do together.
But we'll look at trying to do the whole thing at once and
if we can come up with a solid definition and a good methodology and it
makes sense and it can work, perhaps there's a way to do this all at
once. But practically, I don't see how we can do it and I'm not sure
how anybody at a facility could do it either. I mean, it would be
changing the whole rule overnight.
So I see that this phased-in approach seems to be the most
practical and I'm not sure exactly how that will work out at this point.
MR. HOLAHAN: Perhaps we should be identifying an alpha list
of most important components.
MR. REED: Okay. Then the final task, as we see it --
DR. APOSTOLAKIS: The alpha factor model, though, conflict.
MR. REED: We'll have to get the ACRS' advice on the new
terminology, we can see that right now.
But the last task then is developing some requirements that
would control the method or describe the method for identifying
risk-informed systems, structures and components and some ideas would be
this could be in a reg guide or could be in the appendix to Part 50.
Right now, those are the kind of ideas, and it would include elements
like PRA quality, for example, risk considerations, scope, quality,
detail in the PRA. You can go on and on about PRAs. You guys are very
familiar with that, of course.
What part of the old deterministic defense-in-depth part do
we need? This goes back, unfortunately, to some of the conversations
that started this meeting off, and this is risk-informing. So we need
to try to blend two things together here and that's what that refers to.
Of course, we'll address all modes of operation. We've
learned in the last ten years, I guess, that shutdowns, for example, can
be pretty important risks. As a matter of fact, just as an aside here,
when you look at all modes of operation, you may find things coming onto
the Q list that weren't there before. There are shutdown modes, as a
matter of fact, when you apply this methodology.
Since this is a new thing, obviously, we would like feedback
or some sort of strategy that measures how well it's working and if it's
gone out of kilter, to -- if you move too many things out of the, quote,
special treatment category, to have some sort of mechanism to bring it
back in if you see the overall risk of the plant increasing from
whatever measurement you might have.
And if you do this right, I think, at least conceptually, is
my thought, is if you build this into Part 50, perhaps you could do this
without licensees coming in for prior NRC review and approval, and that
goes along the lines that we don't want to burden industry and we don't
want to burden the staff with all these guys coming in here.
So if we could do this right, perhaps it will work as far as
the most efficient way to implement. So that's what this says, the last
That's what we see right now as our tasks for option two and
we're going to try to build these into a rulemaking plan.
DR. APOSTOLAKIS: Before we proceed to option three, though,
I think we should think about the schedule again. The industry was
supposed to go first and there are some visitors from out of town. Do
you have any planes to catch later? So there is no reason for us to
interrupt this and give the industry time to talk, or shall we continue
with the staff's presentation until they're done. Is that okay?
MR. ROSEN: That's all right with us.
DR. APOSTOLAKIS: All right.
MR. CUNNINGHAM: We'll turn back now to option three and
talk a little bit about what we have in mind for performing that.
Basically, there's two phases that we see in the option three work.
First is the study phase that we talked about in SECY-98-300. The
second phase then is the development of the technical basis and actual
In the first phase, we've got four tasks identified.
Basically, one, we want to look at -- first, identify some alternative
approaches for making Part 50 more risk-informed.
DR. APOSTOLAKIS: Can we stop calling this -- saying for
including quantitative risk information in Part 50? Because we seem to
be using this more risk-informed expression all the time. What we're
really talking about is using quantitative risk information, is it not?
MR. KING: Yes, but it's not clear that it's going to be in
Part 50, that those numbers will show up in Part 50. They may show up
on a reg guide somewhere.
DR. APOSTOLAKIS: Yes, but you're still using quantitative
MR. KING: Yes.
DR. APOSTOLAKIS: Those PRAs.
MR. KING: Yes.
MR. CUNNINGHAM: Yes.
MR. HOLAHAN: You're using the information.
DR. APOSTOLAKIS: Instead of saying more risk-informed, it's
already sent the message.
DR. WALLIS: It's one of those opaque expressions.
DR. APOSTOLAKIS: Or just using PRA.
DR. WALLIS: That makes sense.
DR. APOSTOLAKIS: That would be simpler even, because you're
right, you're not always using the numbers. You may be using other
concepts from PRA, not necessarily the numbers. I think so far, in
several letters, there is quantitative risk information. So including
quantitative risk information.
DR. WALLIS: The risk-informed means using quantitative risk
information and to make better decisions.
DR. APOSTOLAKIS: No, because it's already risk-informed.
But when they designed it the first time around, they did have some
qualitative notion of risk.
DR. WALLIS: But making it more risk-informed, the measure
of success is how risk-informed it is. So you have some sort of meter
of how risk-informed is it. I don't quite understand what the measure
of success here is.
The more PRA we have in the regulations, the better, is that
what it means? Making them more risk-informed? When you come back and
say it's more risk-informed, what would be the your measure of it being
more risk-informed? This is the key thing, isn't it? This is what
you're trying to do. Or is it to try to see what PRA can teach you,
which might make Part 50 better on some other basis?
MR. CUNNINGHAM: Again, going back to what we talked about
originally, what we're trying to do is get to the issue of making
requirements within Part 50 commensurate with their safety significance.
DR. WALLIS: The assumption seems to be if you put more PRA
into Part 50, that you will be somehow better.
MR. CUNNINGHAM: Yes.
DR. WALLIS: So let's inject it everywhere with PRA and be
MR. CUNNINGHAM: That's a good point.
DR. WALLIS: Is there some other ulterior measure of success
other than how much PRA is in there?
MR. HOLAHAN: Yes, I think there is. Sure. The clarity and
efficiency of the regulations, how easy they are to implement, how well
they are understood, how inspectable they are.
I mean, there are other qualities to the regulations.
DR. APOSTOLAKIS: Not all of these are improved by the use
MR. HOLAHAN: Not all of these are improved. But we have to
find some optimum level.
DR. WALLIS: So you have a measure of clarity and you say
here's use of PRA, here's not use of PRA, with the PRA, it's clearer,
therefore, it's a better criterion then.
MR. HOLAHAN: I think that's part of it. Sure.
DR. WALLIS: I think eventually you're going to have to say
we've got some measure of public safety, got some measure of
cost-benefit, and on the cost-benefit basis, this is better than that.
MR. NEWBERRY: I think there is general agreement with those
MR. HOLAHAN: If you go back to the original PRA policy
statement in 1995, it identified three objectives and we currently deal
with another objective of making better safety decisions, and I think
that, in my mind, that's comparable to using more risk information in
the decisions, being more effective and efficient in the use of staff
time and resources, reducing unnecessary burden on the licensees.
In our current regulatory scheme, we're also concerned about
maintaining or improving public confidence by having a more transparent
and understandable regulatory process.
So if you lay those things out, not every one of those is
increased by putting the maximum amount of risk information in every
DR. WALLIS: So somewhere along the line, then, you're going
to make judgments about how much --
MR. HOLAHAN: Yes.
DR. WALLIS: -- how much PRA is enough.
MR. HOLAHAN: Yes, of course, we are. Yes. Now, I think
what you're pushing us --
DR. WALLIS: How safe is safe enough, how risk informed is
risk informed enough.
MR. HOLAHAN: Yes, of course. Wherever we stop we will
think is the right place. You're going to help us find it. That's the
way it works.
DR. WALLIS: I think you have to have a rationale that is
deciding what you're getting.
DR. APOSTOLAKIS: I don't know what how much PRA is enough
means, by the way. I mean, you don't mean that some pieces of the PRA
you will not use.
MR. HOLAHAN: You might not model things at a level of
detail that are wasteful of your efforts. People do truncations now.
Why? Because it's not really improving your understanding commensurate
with what it costs.
DR. WALLIS: Eventually, it's risk-benefit.
MR. HOLAHAN: And you make some decisions about being
conservative in some of the criteria because it's not worth doing all
the analysis to find out the answer much better.
I think the only thing that, Dr. Wallis, you're pushing us
to that we really weren't doing is to think more carefully and maybe
quantitatively about each of these measures, which subjectively we're
saying we're trying to achieve these three or four objectives. In some
decision theory, you could put numerical values on those. You could say
which of our options go how far in each of those.
You could weight each of those in how important they are and
score them in a way. We haven't been doing that in that level. But the
thinking is in there.
DR. WALLIS: I guess what I'm groping for is more the public
perception and the public value that -- it's not clear to a member of
the public that making regulations more risk-informed is good, bad or
indifferent. It's just that somewhere along the line, there has got to
be something in terms of measures which mean something to the public, it
seems to me.
You have a bit of sales task.
MR. HOLAHAN: I understand. That's very difficult to do.
The best explanation I can give them is that we are working towards
making sure that the staff and the licensees are focusing attention on
those things that the risk analysis says is most important. A step in
the right direction.
DR. APOSTOLAKIS: And, again, there isn't such a thing as
the public. I mean, there are many publics, and a good part of the
publics think that using quantitative risk information improves the
So anyway, the message is very clear that perhaps we have to
do a little better job telling people why this is better. And I think
the criteria that you mentioned, Gary, are pretty good, clarity,
consistency and so on. It's just that you haven't really thought about
them enough to put them on a viewgraph.
MR. CUNNINGHAM: We do in a few pages, but we'll come back
in a few pages to that.
DR. APOSTOLAKIS: Okay. That's why I say you haven't yet.
MR. CUNNINGHAM: Yes, correct. We haven't yet.
MR. HOLAHAN: We didn't think enough to put them on the
first viewgraph. We put them on the 20th viewgraph.
DR. APOSTOLAKIS: Okay. So what's the message here on
three? Did we exhaust three -- I mean 16?
MR. CUNNINGHAM: On 16, we're starting to say we've first
got a set of alternatives and we want to evaluate those alternatives
using some criteria that we'll get to. Then we want to look and we're
kind of presuming that that's going to come into some sort of a
screening of Part 50 for the regulations where it's most important to
change. Those regulations that are most important to change, we're
going to revise them, we're going to make recommendations to the
Commission in the first phase as to what regulations to go after.
DR. APOSTOLAKIS: So far, you have not really touched the
content of any of the regulations. So it's all scope.
MR. CUNNINGHAM: Option two deals with the scope. Three is
where we would get into the content.
DR. APOSTOLAKIS: But even in three, phase one.
MR. CUNNINGHAM: Phase one is the study phase, which the
Commission approved, to say, now, which ones should we -- in which
regulation should we change the content and how should we change that
DR. APOSTOLAKIS: You're going to talk about the timeframes
somewhere? Is this is in six months, a year?
MR. CUNNINGHAM: We'll come back. Yes, we're coming back.
DR. POWERS: Let me, before you turn pages and so we don't
go through the viewgraphs too quickly, understand the objectives here
and particularly 1.2 and 1.3.
Is the objective to be complete or to get an adequate
MR. CUNNINGHAM: An adequate sample of regulations or --
DR. POWERS: Requirements for revision. Candidate
requirements for revision. Are you trying to get them all or are you
trying to get some large number or some tractable number?
MR. CUNNINGHAM: Could we come back to that in a couple --
we touch on that in a few slides down.
DR. POWERS: I'm willing to come back and I bet you can even
anticipate what question I'm going to ask.
MR. CUNNINGHAM: Okay. Tom will be covering that slide, so
DR. POWERS: I bet he can anticipate what question I ask.
MR. HOLAHAN: Can I just throw something in? From the point
of view of the office that has to turn these into regulations and
realize that it's not easy to do that, one of the most important things
is that there is a screening process and a study approach that
identifies -- that prioritizes them, not just that collects the 50 that
you might want to do, because that could take you 100 years.
The point is what -- we need to identify the ones that are
most important to do, that have the most value in terms of what criteria
to set up, so that we can do those early. So I'm not so concerned about
the completeness of capturing them as I am of capturing the most
DR. POWERS: I am not sure that that's an easier task to
identify the most important ones versus being complete.
MR. KING: I think there has to be a cutoff here. You could
go through and probably make a lot of modifications that aren't going to
make much difference in terms of reducing unnecessary burden or
clarifying and simplifying things, and we have to find some way to not
put those on the list, to really pick out the ones that are going to
represent some substantial change, either improving safety, reducing
burden, or clearing up something that is not very clear now.
We've got some proposed thoughts as to what we would look at
to do that, but trying to capture everything, no. Trying to figure out
which of the ones we want to do first, yes, and then it will be some
judgment as to where we stop.
DR. WALLIS: You seem to be assuming that making things more
risk-informed will necessarily be better, but it will reduce burden.
Won't there be some instances where making it more risk-informed makes
some things worse? It may require more effort from the licensee, for
MR. CUNNINGHAM: Yes.
DR. WALLIS: So you're going to have to trade off these
various things. It's not as if everything is rosy. You have to have
some way of weighing them against each other.
MR. CUNNINGHAM: Yes. Slide 17, you asked about a schedule.
At this point, the date, the only date that's set is that we owe a plan
for phase one, a plan for the study by October to the Commission. In
that plan is where we will lay out the dates that go beyond that of when
we would finish the study and that sort of thing, and at this point, I
don't have a sense of whether we're talking about six months, a year or
what in terms to get there.
So the first phase one product is the plan for the phase
one, which is October. Then the second part of phase one will be
recommendations to the Commission on what regulations to change and how
to go about it.
The Commission has told us if we come along and see a
regulation right now that needs changing, go ahead and tell us, don't
wait for the end of the study. Something with 50.59 may be an example
of something like that.
DR. APOSTOLAKIS: But if this is scheduled TBD, do you have
some idea when that will --
MR. CUNNINGHAM: I don't have a good sense for it now.
DR. APOSTOLAKIS: Six months, eight months? It's on the
order of months?
MR. KING: I think we're probably talking one to two years.
That's my rough guess.
DR. APOSTOLAKIS: Two years?
MR. KING: By the time we go through and take a look at the
technical requirements, try and assess do they have excessive margin, do
they have things in there that aren't very risk-significant, and then
come back and look at what's the value impact of making a change,
package that all up and send it to the Commission, by the time you do
that, march through Part 50 and get all of that done, that's not a
DR. APOSTOLAKIS: The Commission has told you, as Mark just
reminded us, that if you identify something that can be done, come to us
MR. KING: Yes. And I think what Mark is saying is if one
of these is a really big ticket item, it's going to really reduce a lot
of unnecessary burden or is really going to plug a hole in the
regulations. Yes, that kind of thing, we're not going to wait till the
end and put that in the final paper.
DR. APOSTOLAKIS: But according to this, then, you will not
work on the contents of any of these rules before two years. That's
stretching it out.
MR. KING: No. I think according to this, we send the paper
up October this year that says this is how we're going to approach it,
this is the schedule, then we start doing that. Some big ticket items
may pop out of that that we would send to NRR at that time and so they
could get started on rulemaking.
But by the time we're all done and completed our study, I
don't think we're talking months. I think we're probably talking one to
We're going to be coming back to you periodically, giving
you an update on what's coming out of the study.
DR. APOSTOLAKIS: What is stopping you from deciding now to
risk-inform 50.59? I don't understand. What's stopping you? I mean,
it's certainly something that's of great interest. Already there have
been some attempts. Why do you have to go through this to decide that
it's a big ticket item?
MR. KING: As Mark said, that may be one that we say we've
done enough to know we want to risk-inform this and hand the ball off to
NRR and they'll start a rulemaking. We may decide that.
DR. APOSTOLAKIS: Why don't you do it now? I mean, is it
just because you don't have time to think about it?
MR. KING: No. NRR has been thinking about it. You saw a
couple of viewgraphs they talked about it.
DR. POWERS: The impression was already on their list.
DR. APOSTOLAKIS: No.
DR. POWERS: Wasn't it on your list as a special task?
DR. APOSTOLAKIS: No, that's the scope.
MR. REED: Just the scope.
DR. APOSTOLAKIS: Not the content.
MR. REED: Our discussions are turning into content real
quickly and like I said, if we come up with a clever way to address the
scope, we certainly will. But these guys in -- maybe you can think
about it, already a lot of thought has gone into it, but I haven't heard
anybody come up with anything I've read so far that says there's a good
way to fix this and make it risk-informed.
And let's say you do it, oh, no, no, no, just let me finish
here a second -- if you do it and do make it risk-informed, then the
question is what are you going to apply it to. You're going to apply it
to a different suite of events in Chapter 15, obviously, right? The
ones that are clearly important.
That means they have to be in the licensing basis of the
plant. That means that some utility would have to take those and stick
them into their licensing basis and I'm not aware of too many licensees
that are going to willingly do that.
I mean, perhaps there are, but I think they'd be out of
their minds. I wouldn't do it if I was a licensee.
That's why I'm saying 50.59 is a change control rule that
works on the old way of doing things, you know, the fission product
barriers and Chapter 15 has analyzed for the four categories of events
and bound the criteria for each event. You guys have heard that over
and over again, it works well and basically it maintains that structure,
Now, you go to risk-informed, it doesn't really work. It's
apples and oranges. It doesn't work on Chapter 15. It works on --
DR. APOSTOLAKIS: I guess some of us are not convinced that
it works. A lot of people are saying it works. It used to work before
the attorneys said that zero means zero.
MR. REED: I agree with you 100 percent.
DR. APOSTOLAKIS: Now we don't know that it works.
MR. REED: I think it works --
MR. CUNNINGHAM: This is all presuming that that nuance of
zero or minimal is taken care of.
DR. APOSTOLAKIS: That's a big assumption you have there.
MR. CUNNINGHAM: That's option one. They were told, option
one, the Commission said proceed and finish 50.59 rulemaking on that
minimal versus zero issue.
MR. REED: I don't know. Conceptually, what I'm saying is I
think you almost need like a rule, a risk-informed change control rule
that would work on a risk-informed set of stuff, and that just isn't
Now, as you go to risk-informing Part 50, stuff will be
pulled into your licensing basis. Now it starts to get a mixture of
stuff in there, the old stuff and the new stuff, and now it starts to
become more important. When you do the stuff that's risk-informed, it
doesn't really fit into Chapter 15, how do you control changes to that.
Then it becomes more important. That's what my thinking is
on it right now, but that's not a lot of thought, frankly. But I'm all
ears, if somebody's got a way to do it, then we'll do it.
MR. NEWBERRY: In the context of this discussion, Research
staff and NRR staff both have 50.59 and option two and option three and
what's happening is it's really being looked at together to see what we
should be doing on 50.59. It's a separate task right now.
I think we're calling it a 2(b). Isn't that what we're
MR. CUNNINGHAM: A two and a half or a option two and a half
or something like that. We recognize that 50.59 is one of those ones
that we may have to jump on quickly, but it's not -- as Tim was saying,
it's not a straightforward change.
DR. POWERS: I'm sitting here thinking, gee, Tom sure is
optimistic to get this done in two years. So I said since obviously
some people think that's way too long, I was thinking about why do I
think it's so optimistic to believe it's two years.
I think it's something that you said. You're going to go
through and you're going to look at these things, the existing
regulations, and you're going to say, gee, is this in or out or change
or not change and all that.
It seems to me there is another aspect, when you think about
risk-informing the regulations, that there must surely be regulations
that one would have created, had been from the beginning, if one were
risk-informed, that one did not create when one was more of a
deterministic frame of mind.
And where does that come in into the schedule that you've
laid out? That seems like an extraordinarily difficult job to me to
identify a regulation that should exist, but doesn't exist now.
MR. KING: It's part of the study. For example, maybe a
requirement on feed-and-bleed for PWRs, which is important in a number
of PWRs, that may pop out of our assessment as something that would be
in a risk-informed structure or would be some sort of requirement and so
on, what should those requirements be, would they pass the backfit test.
We need to develop enough technical basis so NRR could
proceed confidently with a rulemaking to put that in. That takes time.
50.46, ECCS rule, I mean, that's probably a likely candidate
to be risk-informed. There's a lot of stuff in there and it has a lot
of implications for other parts of the regulations.
I envision, in the option three study, we'll be going in,
we'll be looking is the design basis accident, the large break LOCA, the
right one; if it's not the right one, what is the right one. The
methods that are used, they've got a number of conservatisms in them.
So those, to me at least, seem excessive.
If we want to take those out or trim those down, what do we
trim them down to? The acceptance criteria, 2200 degrees, 17 percent
oxidation limit, are they the right acceptance criteria in a
risk-informed arena? What's the -- where else does this -- would a
change affect the regulations? What's the value impact of this kind of
There's a number of things to look at and they're not
simple. That's why I'm -- a six-month study is, to me, not in the cards
in option three.
DR. WALLIS: Are you just tweaking existing regulations to
put in more risk information or are you saying if we wrote ECCS rules
today in light of risk information, what would they look like? That
would be a revolutionary change. You don't really need to depend on the
past. You're going to take a fresh look at this whole problem.
MR. KING: I think it's more the latter.
DR. WALLIS: Is that what you're going to do?
MR. KING: It's more the latter.
DR. WALLIS: More the latter?
MR. KING: Do you keep the concept of design basis
accidents? I mean, that's, to me, a policy question.
DR. WALLIS: I would be inclined not to keep anything until
we were driven to it by our new logic and we found, gee whiz, those guys
were smart after all, maybe. Don't start from their basis and tweak it.
DR. APOSTOLAKIS: The reason why I asked the question, Tom,
is that I really don't want us to find ourselves a year and a half from
now still debating which rule to risk inform. The Commission clearly
said that you should pursue this in an aggressive way. So perhaps you
should be a little bit more specific that you will have some first
results half a year after October or something like that. Surely you
can evaluate one or two potential candidates over that period.
MR. KING: I would hope. We're going to do an initial
screening to say, okay, these are the ones we really want to focus on.
I would hope that screening could be done very quickly.
DR. APOSTOLAKIS: Do you plan to interact with stakeholders
MR. KING: Yes, we're going to have to interact with
DR. APOSTOLAKIS: I propose that we take a break. We'll
reconvene at 3:10.
DR. APOSTOLAKIS: We are back in session. Let us continue
with option three discussion. So we are on slide 17.
MR. CUNNINGHAM: We were talking before the break about the
schedules and things and maybe we can go on.
DR. APOSTOLAKIS: Okay. I think we're done with this.
MR. CUNNINGHAM: One last point, I guess, on schedules and
things. Just for what it's worth, we were talking about rules coming,
potential rule changes coming in advance of the completion of the phase
I should note that tomorrow we'll be back, some of us will
be back to talk to you about possible changes to the PTS regulations.
It didn't come about as a result of all this work on risk-informing Part
50, but you'll see a flavor of the changes that we're talking about in
PTS regulations that have this kind of a risk-informed flavor to it.
We'll come back to that tomorrow.
In a sense, it's an early model perhaps of how you would
change some of the regulations. But we'll talk about that tomorrow.
Let's go to slide 18. We talked about this earlier, that
within phase two, there are two parts, if you will. One is once we've
gotten approval from the Commission to change certain recommendations,
there is a technical analysis part and then there's a rulemaking part, a
technical analysis or developing the technical basis for the rule
change, we would have the lead for it in Research. The rulemaking lead
would be in NRR.
So these things would be in Research for a while and then go
back to NRR for the rule change.
So that's probably the key point on slide 18. So we'll
switch then and go over to some more discussion on some of the details
of option three, and Tom is going to pick it up.
MR. KING: Slide 19 is probably the one we should put up
front for Professor Wallis, who is not here right now. Actually, slide
20 is the one that he would be interested in.
Slide 19 is -- when you go into risk-informed Part 50,
there's three basic options. You can sort of start with a clean sheet
of paper and rewrite the regulations, the ones that would be affected by
risk insights, or you can go through the existing set of regulations and
pick out those things that, based upon risk insights, look like they're
excessive or look like there's gaps and plug the gaps and change -- make
it more realistic in those areas, and that would include not only the
regulations, but looking at their subsidiary reg guides, SRPs,
definition of design basis accidents, the analysis methods, assumptions,
acceptance criteria, the whole range of things that flow under a
regulation to actually implement it, and none of those are off limits in
Or the third way, which is probably very simple in terms of
rulemaking, and that is go and pick out those rules where we feel risk
insights could make a difference and allow the licensees to go propose
some alternative, like is allowed now in the regulations for IST.
The licensee can propose an acceptable alternative, the
words are right in the regulation, we could do that in a number of
Of course, if we did that, we'd probably have to figure out,
if we got a bunch of applications, how would we review them. So we'd
have to start thinking about reg guides and SRPs anyway to define how we
would review and find acceptable such a submittal.
But in terms of the actual changes to the regulations, it
would be pretty simple.
So in the paper that goes up in October, we will select one
of these or recommend one of these as the favored approach that we plan
to proceed on and hopefully the Commission will agree with it.
How are going to review those alternatives? Slide 20 sort
of lays out the criteria that we would use in taking a look at those,
but these same criteria could probably also apply across the board when
we go in and actually make changes to Part 50, how are we going to
decide whether the change is actually an improvement or not. There's a
number of things we could look at.
Probably a lot of this stuff you'd look at qualitatively,
although some things, like safety impact, you probably could come up
with some quantitative judgment as to whether things have improved,
stayed the same or gotten worse.
And these things sort of flow out of the three major
purposes that were stated in the objectives in the PRA policy statement.
We'd want to look at safety impact, we'd want to look does it reduce
unnecessary burden, does it improve clarity and consistency, the
resources, is it practical for NRC and the licensee to implement from a
resource standpoint, does it take excessive time, and how would you
inspect and how would you enforce such a regulation.
So these are the kinds of things that we will be looking at
as part of making any decision to go propose a change to the
On to slide 21 --
DR. WALLIS: On the unnecessary burden impact, as you said
before, it may turn out that you examine some of these regulations and
there may be more burden. It goes both ways.
MR. KING: You're right, you're right.
DR. WALLIS: Thank you. I just wanted to confirm that.
MR. KING: A change will mean somebody has to do something
different, and that has a burden with it. That's true.
DR. KRESS: Can you really use these criteria to evaluate
your first alternative very well? It seems like a difficult thing to
MR. KING: Resources required, I think, would be a --
DR. KRESS: That would be the one that would hit you hard.
MR. KING: Get a high score or low score, depending on which
way you -- what's favorable. Time required would certainly be impacted
by the first alternative.
DR. KRESS: It would be tough to do some of the other
things, like safety impact and burden impact, until you already had the
new version in place or some idea of what it was.
MR. KING: Yes. I mean, my own personal opinion is we'll
probably go with the second alternative here, which is to go in and look
at the existing set of regulations, identify those that are really
candidates for change and go fix those.
There's probably a lot of stuff in there that's okay as is.
DR. KRESS: Is there a possibility of a bit of compromise
between those two? Like if I were going to start from a blank sheet,
what would I want to get accomplished by a new Part 50? Maybe that's
not as big a deal as actually putting together a new version to make
those accomplishments, but then use those to guide what you would do
with the second alternative.
MR. KING: That's a good idea. Use it as a template to say
DR. KRESS: A template to say this is what I want to get
accomplished and you would have things like how you deal with
defense-in-depth in that and how you deal with uncertainties in the PRA
and things like that. You'd say we want to get a certain level --
accomplish a certain level of risk status probably would be one of the
things we'd want to accomplish.
MR. KING: Yes. Almost what you're talking about is a
comprehensive set of objectives and framework if we were going to
actually do that and then say, okay, how does that overlay on the
existing Part 50.
DR. KRESS: How does it overlay on the existing one.
DR. APOSTOLAKIS: But didn't the GAO report blame the NRC
that we didn't have a strategy for achieving objectives that we have not
defined in going to risk-informed regulation? Do you think your
approach here will satisfy maybe critics like the GAO or we need to do
something else or we decide not to heed their advice?
MR. KING: Actually, the Chairman responded to GAO and said
we are going to develop a strategy.
DR. APOSTOLAKIS: So the PRA implementation plan will do
MR. KING: The PRA implementation plan will evolve into a
strategy document. But I view the GAO report as at a higher level than
how do we change 50.46. I think they were looking at what areas of
everything we do are we going to get in and risk inform. The NMSS
activities, the --
DR. APOSTOLAKIS: But risk-informing Part 50 --
MR. KING: Research reactors, I mean, decommissioning,
whatever it may be.
DR. APOSTOLAKIS: But risk-informing Part 50 must be at the
same level of that report.
MR. KING: I think an element of the strategy would be
risk-informing Part 50, what do you have to do, what are the goals and
objectives, how do you measure success.
DR. APOSTOLAKIS: Yes, but in choosing, though, to
risk-inform individual or existing regulations, then you are not really
doing what they want us to do. In other words, it seems to me that if
you -- what was it -- there was a -- you had the line here somewhere
that you could start thinking from scratch, right?
MR. KING: Yes, slide 19.
DR. SEALE: That's the clean sheet.
MR. KING: Slide 19.
DR. APOSTOLAKIS: Entirely new alternative. I mean, if you
attack the problem this way, then you can claim legitimately, it seems
to me, that you are responding to criticism of that kind, like the
GAO's. But if you start changing individual regulations and making them
better, then it's questionable whether there is a clear statement of
objectives and strategy to achieve those objectives.
DR. SEALE: Well, it strikes me that you've already opened
the door and received some results of part three. I mean, after all,
1.174 is a document that facilitates the pilots and the pilots or
industry or the applicant derived or proposed versions of the
So in certain specific areas, you've already got that and
what you're talking about now, I guess, will be to add to that an
organized staff approach to other elements using the second proposed
technique, but you're certainly not going to throw the pilot results out
the window and start over on them, I don't think.
MR. KING: No. Gary and I had several discussions with the
GAO folks who wrote the report and I can give you my impression and Gary
can give you his.
But I think what they saw was they saw Reg Guide 1.174, they
saw the specific guides for IST and so forth, and they said, gee, that's
nice, but you're just picking and choosing certain areas; what's the
overall picture for the agency; what other areas are we going to go
into; how are we deciding what those areas are; how are we deciding --
do we have the resources, do we have the training, do we have all these
higher level things that will make sure that stuff gets done.
I think that's where they were coming from in their report
and the strategy evolved in the PRA implementation plan into a strategy
document is going to come at it from that level, because the PRA
implementation plan covers NMSS stuff. It covers enforcement, it covers
a lot of things beyond Part 50.
It's not going to -- I don't think GAO is interested in how
are we going to risk-inform 50.46.
DR. UHRIG: But given the resources that you have available, are you not
somewhat restricted in trying to take the first alternative here? That
could be a massive job and it might be a fruitless job in that a lot of
what you would do here would have no impact at all on industry if they
weren't interested or they perceived no benefits of it. That's the
beauty of coming from the bottom.
As Bob pointed out, they have picked the things that clearly
to them look like it's going to have some benefit.
MR. KING: I agree with you.
DR. UHRIG: And put your resources there and as a result, one hopes that
there is progress being made, albeit with problems along the way, I'm
MR. KING: I agree. I think --
DR. KRESS: Yes, but I don't think the industry would have
these other criteria as first priority that you have. You have this
list of criteria and I think starting from the bottom, you might miss or
at least lose some emphasis on some of those others.
DR. SEALE: You get an incomplete product and that's why the
second element is necessary.
MR. KING: But recall, in the vote sheets in 98-300, there
was one Commissioner who said start with a clean sheet of paper. That
was his view.
DR. KRESS: Yes, I remember that.
MR. KING: So I think we need to address that in going back
to the Commission and say yes, we considered that and here is why we
don't think it's --
DR. APOSTOLAKIS: So there are pros and cons, there is no
question about it.
MR. KING: Exactly.
DR. APOSTOLAKIS: If you started with a clean sheet of
paper, then the benefit will not be apparent for a while.
DR. WALLIS: This move to risk-inform is a step of faith, is
it? I would have thought that in order to justify the direction,
someone must have made some preliminary calculations and an assessment
of these things and said if we did that, gee whiz, we could cut the
burden in half and so on, first estimates indicate that we could get the
same safety at half the cost or something. There has got to be
something like that. It's not just invoking something and hope that it
might work, is it?
MR. HOLAHAN: What we're doing is we're building on the
experience from the activities already accomplished under license
amendments and Reg Guide 1.174. We've seen licensees reducing burden
and, in virtually all the cases, improving safety.
So that's an indication that areas like QA and probably
maintenance rule and a number of these other areas are likely to drop
out of this sort of test with very favorable results.
DR. WALLIS: Just a sense. I think if industry were doing
this, they'd have someone look quickly at it and say we've had some good
success with 1.174, let's see if there is any payoff to extending it to
others; go away and come back in two months and tell me whether there is
a payoff, where we should put our effort.
You seem to be still not there yet, knowing if there is a
payoff or where to go.
MR. KING: And I think we're going to need a lot of input
from industry in terms of certainly the unnecessary burden reduction
piece. They're in the best position to know what's costing them money
and where changes could be made that don't have any impact on risk.
DR. WALLIS: Are they jumping up and down saying you've got
to do it and why are you taking so long?
MR. KING: You're going to hear from them later today. You
can ask them that question.
DR. APOSTOLAKIS: They will be here in a few minutes. In
fact, can you finish in about 15 minutes, Tom?
MR. KING: Sure. Okay. Let's go to slide 21.
The next three slides talk about what I call potential
policy issues. We are still formulating the policy. That's why the word
"potential" is there.
These aren't necessarily limited just to option three. Some
of these could apply to both options two and three, the scope and the
technical requirement part. Ones that we've sort of put down so far in
terms of policy issues, the first one has to do with when we risk-inform
Part 50, what's the basic approach we're going to take in terms of do we
just want to make adjustments around the current risk profile of a
plant, sort of like Reg Guide 1.174 does, tells you how much you can
move around that risk profile, or do we actually want to go in and try
and, on a more absolute basis, put in requirements that would tend to
bring people, all the plants to some common ground safety goals, for
DR. APOSTOLAKIS: Or both.
MR. KING: What?
DR. APOSTOLAKIS: Both?
MR. HOLAHAN: 1.174 does a little bit of both by having
stricter criteria on higher risk plants and generous criteria on lower
DR. APOSTOLAKIS: So you could conceivably have a
combination here, as well.
MR. KING: You could. You could.
DR. KRESS: Are these things you will want ACRS to comment
MR. KING: We're going to get to the schedule. Not today.
DR. KRESS: When you get to a paper.
MR. KING: I'll take any verbal comments you have,
DR. KRESS: But these are policy issues we can be thinking
MR. KING: Yes. But we're not asking for a letter out of
today's meeting. Eventually, we're going to come back with our --
DR. APOSTOLAKIS: But sometimes we write without being
MR. KING: If you want to write one, that's up to you. We
are going to propose to come back to you in September and October and,
at that point, get a letter, where these things will be more refined and
Second policy issue is should we retain the design basis or
should we risk-inform, that would even give more flexibility for
licensees to come in and propose something for their plant. Retaining
design basis doesn't mean retaining today's design basis. It might be
some more risk-informed design basis, a better -- a more realistic
design basis accident, for example, more realistic assumptions.
Burden reduction revisions, should there be some test that
the burden reduction you're getting is really worth it, you're not
giving up a lot of safety. I call it a reverse backfit test; do you
apply the same rules that you would apply on a forward backfit.
The fourth one is sort of a carryover from something that
was said earlier, the selective implementation issue. Remember, the
Commission said it was premature to decide that one. So at some point,
that's going to have to go back to the Commission as a policy issue.
And then the adequate protection question that came up on
our second or third viewgraph, we go in and make all of these changes,
how do we know we still have adequate protection. If we're making
changes that aren't all that great, it's probably not that big a
DR. WALLIS: I would hope that you would, as a result of
risk information, you would be assured, better assured about what you
meant by adequate protection. It's another question of preserving
something. It's knowing that you really have it.
MR. KING: Well, by definition, we have it today. We're
going to go in and make a bunch of changes.
DR. WALLIS: I'm not sure I would agree with that. I'm not
sure whether all these design bases and things, which are very fine and
have worked, really have anything to do with --
MR. KING: If we didn't have it, a plant shouldn't be
DR. WALLIS: In terms of risk, in terms of risk metric.
They're all on the same scale, different planet.
DR. KRESS: They're in the same orbits.
DR. WALLIS: They may have the same effect, but we don't
really know it until we do some risk study.
DR. APOSTOLAKIS: If you tie risk-informing Part 50 to the
Commission's safety goals, then aren't you also answering the adequate
MR. KING: I think you are. I think you are.
MR. HOLAHAN: In fact, we're probably going to have to touch
on the adequate protection definition issue in order to do the
clarification the staff has already -- that would be the first place.
DR. KRESS: When you talk about -- think about policy issue
number one up there. Are you also going to ask yourself the question
whether the current safety goals are the right ones to have?
MR. KING: That's a separate paper that's being worked that
will be coming back to this committee in the fall.
DR. KRESS: But that would impact on this.
MR. KING: Certainly.
DR. APOSTOLAKIS: I still -- I'm sorry.
DR. KRESS: I was going to go down the list of these. On
the second bullet, you might consider a full PRA as the ultimate design
basis approach and any shortening of that, where you have surrogates, is
all right probably, because the PRA is actually a selected set of
accidents, too. But it's not a complete set, and they bend things.
So you might be able to run the concept of design basis,
this dichotomy between the two would be awfully nice to get rid of and
this may be an opportunity to do that. Make them comparable.
MR. KING: That's sort of getting into the chicken-and-egg
question, though, and today's plants are designed and built and
operating and have a PRA.
DR. KRESS: I'm obviously already getting a sort of
backhanded answer to the question of whether you should relate it to the
current design basis or you should start over.
MR. KING: One of the things we have to keep in the backs of
our minds is what does this mean for future designs. Whatever we do on
Part 50, can a future designer pick it up and take advantage of it.
DR. KRESS: Yes. And I don't think you ought to get into
the backfit issue at all here on the third bullet.
DR. APOSTOLAKIS: Explain to me the reverse backfit again.
DR. KRESS: Is the burden reduction you get worth the safety
benefit you gave up, and I think that's a problem that you're going to
-- that's going to give you all sorts of problems, because most of these
things are supposed to improve safety in the first place, and you're not
supposed to be reducing safety.
DR. APOSTOLAKIS: In other words, if you have a delta CDF
DR. KRESS: Yes. Is the burden reduction worth it?
DR. APOSTOLAKIS: He already told us what's acceptable
anyway. So I don't see --
DR. KRESS: Yes. I would probably try to avoid the third
DR. APOSTOLAKIS: The staff has already declared that
ten-to-the-minus-five delta CDF is acceptable, without concern about the
DR. KRESS: And --
DR. APOSTOLAKIS: So how would this be different?
MR. KING: There are some general words in 1.174 that imply
the burden reduction. They just didn't come in and make frivolous
changes upping delta -- making changes to delta CDF without getting some
commensurate unnecessary burden reduction. There are no numbers in
DR. APOSTOLAKIS: Sure, but the licensee would not go
through the trouble of preparing the request if they don't see a benefit
at the end of it. So that's sort of built into it, it seems to me.
They're not going --
MR. KING: Qualitatively.
DR. APOSTOLAKIS: -- request just to change the delta CDF.
MR. KING: Qualitatively, it's built in. There is no
numerical guideline to see how well that's complied with.
DR. APOSTOLAKIS: I think Dr. Kress' suggestion probably --
MR. KING: Okay. We'll discuss it.
DR. APOSTOLAKIS: I didn't expect you to do it right now.
DR. UHRIG: Just one comment on the application of PRA to new plants.
When the advanced neutron source reactor was being designed, PRA was
used extensively in the design in order to minimize the potential risk
associated with that plant. It was very successful. Unfortunately, the
plant was never built, but it was used.
DR. KRESS: Yes. I can't imagine any new plant not doing
DR. APOSTOLAKIS: It's not clear to me, though, that even if
you do that, if you use a PRA, that you still don't need some sort of an
envelope, like a design basis set of accidents. You may use the PRA to
select them, but you will still need something.
MR. HOLAHAN: Personally, what I would like to see
eventually, and I don't know whether this is practical or not, I would
like to see an FSAR with thermal hydraulic type calculations which are
related to the success criteria in the PRA.
DR. KRESS: That would be --
MR. HOLAHAN: I mean, that's the tie-in between equipment
performance and whether it can do the safety functions and the
assumptions that you take from there and put in the PRA.
The problem we have now is that there are different
sequences and they have different criteria and they have different
success. So when you have a success in Chapter 15, it doesn't
necessarily tell you that you have a success in the PRA, or sometimes it
produces something that's a lot more successful than it needs to be in
I think that would be a nice coherent place to be. Whether
it's worth all that effort of rewriting Chapter 15 to get there, I
think, is a very practical question.
We may be able to make a lot of progress without doing that.
That gets to the burden reanalysis type point. But that might, in fact,
get you somewhat close on the 50.59 issue, because then at least your
accident analysis that's in your FSAR has some relationship to the PRAs
and you could have risk-informed tests of 50.59.
DR. BONACA: Then you would have to deal separately with
each of the uncertainties. The difference between Chapter 10 and
Chapter 15 is because of the uncertainties and Chapter 15 attempts to
deal with them by having gross assumptions and conservatisms. So you
would have to -- I think what you're suggesting is attractive, because,
again, in 50.59, the closure of the issue is more how do you then treat
the uncertainties separate from actual modeling.
MR. HOLAHAN: Yes.
DR. BONACA: Any thoughts about that?
MR. HOLAHAN: I think this is not a very well developed
idea. And if you look at Chapter 15, it has a lot of other things in it
which are not really core melt issues.
DR. BONACA: That's right. Yes.
MR. HOLAHAN: DNB ratios and stuff like that, which are
important in some defense-in-depth concept. You don't want everyday
events be challenging to core damage. So there ought to be other
criteria and how you say that those are risk-informed or make sense
isn't at all clear. It's not clear to me.
DR. APOSTOLAKIS: Can we move on?
MR. KING: Okay. Slide 22, potential technical issues.
DR. APOSTOLAKIS: We have already discussed some of them.
MR. KING: Yes. We have already discussed some of them and
metrics and criteria for identifying candidate changes, we'll get to
that on slides 24 and 25. We talked about there are a number of things
you could risk-inform in the regulation to the design basis accident and
Once you decide, yes, that's a candidate for change, then
you've got to ask yourself the question, if I'm going to change it, what
am I going to change it to, what is this new design basis accident. And
there's probably a number of options or things you could -- criteria you
could use to decide where am I going to draw the line, and I just sort
of listed some ideas here. We won't go through all of them.
DR. APOSTOLAKIS: This is a variation of the Commission's
SRM decision recently. Are these issues prematurely before the
MR. KING: We're not asking for a --
DR. APOSTOLAKIS: We'll address them when you come up with
something more specific, because we can talk forever about best estimate
with uncertainty, realistically conservative, conservatively.
DR. SEALE: We have.
DR. APOSTOLAKIS: Twenty-three, Tom.
MR. KING: Twenty-three, process issues, we don't need to
DR. APOSTOLAKIS: And there is certain time pressure. NEI
wants to finish by 5:15. So it seems to me we ought to give them a
chance, and we're getting into areas now that I am sure we will have an
opportunity to discuss again with you.
MR. KING: Okay. I think we can probably even skip 23.
They're just process issues.
DR. APOSTOLAKIS: Let's skip 23. That's beautiful.
MR. KING: Let's talk about 24. How are we going to go in
and look at Part 50 and decide are we going to make some changes or what
are candidates for change, and in doing that, we need to look at what
should be added as well as what do we need to modify or remove.
A couple of assumptions that we think probably would apply
in the screening process. One is we're going to focus on the technical
requirements, although there's probably a couple process things, like
50.59, that will be looked at. For the most part, it's technical
Another ground rule we would view is whatever the
risk-informed scope definition is that comes out of option two, we have
to consider that's going to be applied ultimately across the board to
the things that are being looked at in option three.
So given that, flip the page to 25, how do you -- what are
we going to look at in deciding is a regulation candidate for change.
Well, there's a number of things thought about. We could look at, for
design basis accidents, particularly, whether there should be some
cutoff on the frequency of initiating events. Take a large break LOCA,
for example, if it's beyond -- I just threw a ten-to-the-minus-six per
reactor year frequency in, then the large break wouldn't be the design
basis accident anymore, it would be something that would be more in the
neighborhood of a ten-to-the-minus-six range break.
DR. APOSTOLAKIS: Is this the same thing as saying design
the reactor so that the return period of the initiating events is
greater than one million years? Make sure it's greater than one million
years. And then I guess the licensee will have to decide how to meet
that. That's what you're saying here; that if their large LOCA is a
frequency which is inverse of the return period, you make sure, if you
can make a convincing case that it's greater -- less than
ten-to-the-minus-six per reactor year, then it's not the design basis
accident. Is that what you're saying?
MR. KING: In effect, that's what I'm saying. Same thing
for when you look at a scenario, how far do you go in looking at
multiple failures, for example. We could put some frequency cutoff on
scenarios so that a single failure doesn't make sense anymore when
you're looking at realism, what does make sense.
For deciding does something contribute to the safety or risk
significance of the safety functions, we could use the delta CDF and
delta LERF --
DR. APOSTOLAKIS: These are deltas, they're not --
MR. KING: Those are deltas, yes, small deltas.
DR. APOSTOLAKIS: There is a command set of characters in
WordPerfect that has a delta, a real delta.
MR. KING: Okay. Not these little triangles.
DR. APOSTOLAKIS: These are bullets.
MR. KING: And for shutdown, CDF probably still makes sense,
but LERF doesn't, so we'd have to figure out what to do there.
Going in and looking at excessive margins, how would we do
that. Mark mentioned earlier, to me, we'd have to go in and take a look
at the analysis assumptions, the acceptance criteria, and decide maybe
DR. APOSTOLAKIS: What's really missing, Tom, is what I
think Gary touched upon earlier. How do we make the connection between
the physical quantities that we can measure and calculate the failure
rates? As we do that, you can't really use PRA design. You're not
measuring failure rates, right? You are measuring flow rates, you are
measuring temperatures, you're measuring things like that.
And then all of a sudden, the PRA guys come and say, well,
the failure rate of this component is this, under certain conditions,
and that connection is not clear. I mean, the failure rate comes
largely from operating experience. So there is a disconnect there, and
I see that as an impediment.
In other words --
MR. KING: You've got to look at both.
DR. APOSTOLAKIS: -- we don't have good reliability physics
models. I would put it that way. Because the reliability physics is
really that relates the failure of a component, the probability of
failure with, as I said, physical quantities, pressures, temperatures,
flow rates and so on. And our failure rates really do not come from
these kinds of calculations.
MR. KING: No.
DR. APOSTOLAKIS: The success criteria use, you know, do you
need to steam generators or one steam generator under these conditions,
but then you go back to failure rate calculations.
MR. KING: Yes, but I guess I don't -- I think you can
accomplish a lot without getting into --
DR. APOSTOLAKIS: You can accomplish a lot, I am not saying
it's useless, but I think in order to go all the way to a rational use
of PRA in the design, you really need to make that connection, in my
DR. WALLIS: In something like ECCS, it's not failure rate
of a component, but failure rate of the whole thing, because of
uncertainties in the critical flow and the flooding model here and all
these interactions of the components. There is some probability that
you might damage the core even though deterministically it's --
MR. KING: That's right, exactly.
DR. APOSTOLAKIS: Right.
DR. WALLIS: That's very difficult to assess.
DR. APOSTOLAKIS: Yes, but that's the reliability physics.
MR. KING: That's right. But one thing we could do is go
in, do best estimate analysis on a LOCA event today and compare that to
the design basis type analysis and see what kind of --
DR. WALLIS: Then you'd have to be more specific about what
you mean by best estimate.
MR. KING: Yes.
DR. WALLIS: You've got to quantify the uncertainties and so
MR. KING: That's right, that's right, and that's the kind
of sort scoping analysis we're proposing to do under this option three
study. That's why, to me, it's not a six-month job.
DR. WALLIS: You might actually have to hire a thermal
MR. KING: We may. That's scary.
The defense-in-depth, I viewed using RAW and Fussel-Vesley
as a way to determine how much do components or systems or functions
contribute to -- well, components or systems contribute to the functions
we're trying to perform and maybe get some measure of the contribution
DR. APOSTOLAKIS: That's the first time I see RAW and
Fussel-Vesley being used as measures of defense-in-depth. I'm not sure
that I can -- I'm not sure. Especially the Fussel-Vesley. The
Fussel-Vesley is probably not a good measure. I don't know about RAW,
I'd have to think about that. I mean, Vesley himself has commented that
if you have a well balanced design, where every single component
contributes equally to the final result, then all of them have the same
Fussel-Vesley measure. But that doesn't tell you anything about how
much redundancy you have and so on, because they all contribute equally,
MR. KING: Okay.
DR. APOSTOLAKIS: So I'm not sure that that really belongs
DR. WALLIS: However, I'm happy to see some measure of
defense-in-depth, rather than just saying it's --
DR. APOSTOLAKIS: Which, in my mind, the measure of
defense-in-depth is the PRA results. That's exactly what it is.
DR. KRESS: That is a measure of how good the
defense-in-depth does in achieving your risk objectives, but I think
defense-in-depth is there because of uncertainties in the --
DR. APOSTOLAKIS: That's why the PRA is a measure.
DR. KRESS: Yes. And the PRA is a measure of
defense-in-depth if you also include the uncertainties in there and if
you say -- and if you look for values of uncertainty that you -- if you
can relate those to defense-in-depth somewhat, you have to tie those
DR. APOSTOLAKIS: Yes. I mean, when I say PRA, obviously, I
mean including the uncertainties. But also it takes a sophisticated
analyst who would make a distinction, for example, between just
redundancy and diversity. In many PRAs, you really don't do that.
MR. KING: But you could?
DR. APOSTOLAKIS: What?
MR. KING: You could.
DR. APOSTOLAKIS: You could, yes. Then it's a measure.
Then it's a measure.
MR. KING: And for a new --
DR. APOSTOLAKIS: If you don't have sufficient
defense-in-depth, you will not get the numbers and the curves that you
want. The curves, not the numbers.
DR. KRESS: But I'm looking to get both the acceptable risk
and acceptable uncertainty, and that's what I was trying to say.
DR. APOSTOLAKIS: Acceptable risk can be added to it, but
the PRA itself, if done properly, is a measure of defense-in-depth. I
mean, you wouldn't get the ten-to-the-minus-fours and fives that you get
now for CDF --
DR. KRESS: Unless you add something.
DR. APOSTOLAKIS: -- unless you had tremendous
defense-in-depth in the -- especially in the nuclear business, where you
have all these multiple reviews and people are trying to figure out ways
that you are wrong or you don't realize it.
MR. KING: But there may be some things in there that have
the defense-in-depth label on them, which, if you really look, if you
didn't have them, they aren't going to make much difference.
DR. APOSTOLAKIS: Sure.
MR. KING: That's what I was trying to get at with this
DR. APOSTOLAKIS: Sure. That's certainly true.
DR. KRESS: RAW would certainly be a measure of that.
DR. APOSTOLAKIS: Yes.
MR. KING: Okay. And then cost-benefit we already talked
The last slide, schedule. And this is the schedule between
now and the end of October. We've already had a couple of public
meetings, the dates are shown here, where NEI and some of the utility
folks have come and, in general, talked about what we're doing, although
not in the depth we talked about it today, and then, in general, talked
about what they're doing.
We have another one scheduled for July 28, and we need to
schedule some others over the summer.
In terms of interacting with the committee, we would expect
or are planning to draft up the Commission paper for both option two and
option three and try and come back to the subcommittee in September to
discuss that, and full committee in October, at which point we'll be
asking for a letter.
DR. APOSTOLAKIS: So September what, the third week or so,
MR. KING: Fourth week.
DR. APOSTOLAKIS: The first two weeks are out.
MR. MARKLEY: We can decide that here, if you want, or we
can work it out later.
DR. APOSTOLAKIS: But I want to get an idea, what you have
in mind. The October meeting is the first week of October?
MR. BARTON: Yes. It's real early, yes. First, second,
third of October or something like that.
MR. MARKLEY: It's actually September 31st and --
MR. BARTON: September 30th, the first and second.
DR. APOSTOLAKIS: September 32nd.
DR. SEALE: Only in a new millennium.
DR. APOSTOLAKIS: When do you think your document will be
ready for us to review? If you want a subcommittee meeting, you have in
mind that you will give it to us at least a week in advance.
MR. KING: I'd like to get you something at least a week in
advance. So the later in September --
DR. APOSTOLAKIS: It's got to be something at least a week
MR. KING: The later in September the better.
DR. APOSTOLAKIS: Because the schedule after that is really
tight. I mean, you cannot come back with another document after the
MR. NEWBERRY: Let me comment on that. I think Tom might
have been thinking just option three.
DR. APOSTOLAKIS: Yes.
MR. NEWBERRY: I think this is going to be a subject that I
would propose we have some dialogue with the committee, because I don't
know the answer to all the questions, especially when we hear the
industry's presentation. They're coming in with some significant
information in the fall, and maybe NEI is going to talk about that, and
I don't think the staff has figured out how to get from here to there
with a product to the Commission when we haven't yet seen what they're
doing, although we're going to meeting with them on it.
So I think we'd need to think about that.
DR. APOSTOLAKIS: Let's put it this way. If we are going to
have a subcommittee meeting, it would be nice to have a document to
review that is in a reasonable shape a week in advance.
MR. NEWBERRY: Absolutely.
DR. APOSTOLAKIS: Otherwise, we move the subcommittee
meeting further into the future.
MR. NEWBERRY: Absolutely.
DR. APOSTOLAKIS: Okay. Shall we hear from the industry
now? I'm sorry.
DR. WALLIS: You had an organizational chart on page nine
here, and I'm trying to figure out, are you just working on plans to do
things or are these -- does this organization exist and are these people
now working on the problem?
MR. NEWBERRY: Yes.
DR. WALLIS: What stage are we at here?
DR. NEWBERRY: Yes. The SRM was dated, what, June.
DR. WALLIS: So these people are already working all these
problems. So you should not really show review of plans, basically.
MR. NEWBERRY: They've been working on this for about two
DR. WALLIS: You should show some progress, though. It may
well be you've redefined the problem, they've learned some things.
MR. NEWBERRY: That's right.
DR. WALLIS: So I'm not looking just for another plan. I
would be looking for some results of this rather large team doing work.
MR. NEWBERRY: In the fall, you mean.
DR. WALLIS: Yes.
MR. NEWBERRY: Yes. I would hope we would give you more on
DR. WALLIS: I'm glad you clarified that for me.
MR. NEWBERRY: Well, the Commission has asked for a
rulemaking plan. I think we've got to think about what that plan is,
would that plan include a proposed rule or a set of rules.
DR. WALLIS: With all this big team thinking for several
months, you should have made a lot of progress.
MR. NEWBERRY: Absolutely. Yes.
DR. APOSTOLAKIS: I'm glad you agree. Okay. You will be
here during the industry's presentation and the discussion afterwards?
MR. NEWBERRY: Yes.
DR. APOSTOLAKIS: Thank you very much. Very informative
presentation, as usual. Okay. NEI and South Texas Project, I
understand. I guess we can ask you to do the same thing we asked the
staff could do.
Would you put your name on the cover page and your e-mail,
in the future?
MR. GRANTOM: Okay. You're looking at the wrong
DR. APOSTOLAKIS: Obviously I am. These have the names
without e-mail, this is nameless. Name of the presenter, that would be
DR. KRESS: Just assume it's stapled together.
DR. APOSTOLAKIS: Oh, this is NEI. This is different.
MR. ROSEN: I'm going to lead off, George.
DR. APOSTOLAKIS: Oh, I'm sorry. Steve.
MR. ROSEN: My name is Steve Rosen. Rich Grantom, from our
PRA group, is here with me. We're going to talk about implementation
problems and successes at South Texas and one of the pilots that you've
talked about before.
We're pleased to be here, and we've been here before, and
appreciate the opportunity to talk to you.
You will recall that the staff gave us a staff safety
evaluation report back in November of '97, which allowed us to change
our operations quality assurance program to implement graded quality
assurance at South Texas.
Our expectations for ourselves and the staff's expectations
for us, were that we would do it in a professional way and to meet our
expectations, we developed a very robust risk-significance determination
process, so that we would know the risk significance of any component
that we were about to change the quality assurance provisions on, so
that we would know what we were dealing with.
And we're here to tell you about the results of using that
process. We also had some expectations for ourselves in terms of safety
improvements for the plant and in terms of burden and resource
With regard to safety improvements, we knew at the outset
that there were components in the plant that were not considered
safety-related, that had, we thought, risk significance, and there were
many, many components that were considered safety-related that had
little or none. So that was our intuition and now we're here to tell
you, after analyzing some 28,000 components, what we have found in 22
We also had expectations for ourselves and our financial
managers had expectations that this would be a cost-beneficial process
and in the early days, we estimated that we could save about a $1.3
million a year after we had fully implemented this program.
Our estimates have since gone up to where we feel that we
could, if we could fully implement this program, save on the order of $2
million a year.
But there are some impediments, key impediments, and we've
talked about some of them already. They're mainly embodied in the
special treatment requirements and also in the use of Part 50.59.
As you see on this next slide, the staff safety evaluation
did not allow us to use, in any way, any exemptions in the special
treatment requirements. We were stuck with continuing to comply with
the ASME code, with the requirements for environmental qualification,
with the requirements for seismic qualification, even though South Texas
isn't what is often called an aseismic zone, and with the requirements,
for example, for Class 1-E.
So while we were unable to change any of those kinds of
things, we were allowed by the change that we were glad to make to our
operations quality assurance program, we were allowed to change the
quality assurance provisions applied to those components.
DR. UHRIG: Did you ask for special treatment in your application?
MR. ROSEN: No.
DR. UHRIG: You did not.
MR. ROSEN: It was considered, but we didn't do it. It was
discussed with the staff at the time, but we chose not to, because at
the time, it seemed a great step to go as far as we went.
But as you can see, we have had some of our implementation
restricted; specifically, the procurement benefits have not been
And what I mean by that is that if we had a component that
was expensive and we needed to replace it because it wore out, and it
was not risk-significant, we are still required to purchase it with all
the environmental qualification, seismic qualification, et cetera, on
it, even though we have shown that its failure is not risk-significant.
So a little more background, please. SECY-98-300, you've
already talked about it, recommended a phased approach to risk-informed
Part 50. The subsequent SRM authorized the staff to proceed.
Now, the previously approved SDP process, which the staff
granted us authority to use back in November of '97, had, in our view,
anticipated the efforts required to risk-inform Part 50, because it
really did require us to develop a robust risk significance
determination process, which was at the heart of our comprehensive risk
management program and provides the engine under which we use to power
all of our approaches to risk-informed changes.
So how do we go forward? We don't really know for sure and,
fortunately, we're sitting on this side of the table instead of the side
of the table you're sitting on. But here are some thoughts, before we
get to talking to you about what we found.
It seems to us that the scope of Part 50 should be amended
by redefining safety-related, because there is really nothing wrong with
the regulations in Part 50, as long as they're applied to the right
things. So we think it's largely a scope issue and, also, as necessary,
providing selected exemptions to Part 50, and that's what option two of
the SECY and the subsequent SRM are all about.
We think that the process that was approved by the staff
that we use in our graded quality assurance program are risk
significance determination process satisfies the intent of option two.
It is an improved method.
So we come to this set of equalities, which -- what did you
call them earlier today, Rick?
MR. GRANTOM: Basic precepts that are important here and a
fundamental principle here that one needs to accept.
MR. ROSEN: Things equal to the same thing or equal to each
MR. GRANTOM: Yes. This is the transitory law here we're
going to work on here.
MR. ROSEN: I couldn't remember that.
DR. WALLIS: What is the measure of important to safety?
MR. GRANTOM: The measures that we're using here are the
traditional ones that we've talked about here in PRA, core damage
frequency and large release frequency, for the PRA portion of that.
However, the risk significance determination process is a blended
approach in which deterministic engineering information and qualitative
information is also used.
So the actual risk significance determination occurs after
those two sets of information have been merged together. Then there is
a determination on what the risk significance is.
DR. WALLIS: Important means bigger than some --
DR. APOSTOLAKIS: It's really not CDF and LERF. You are
using the risk importance measures. It's not clear really what the
connection is to the CDF and LERF.
MR. GRANTOM: We have two approaches. There is the PRA
information, which uses a -- I will call it a risk ranking approach, in
which we do a whole series of sensitivity studies, that use traditional
figures of merit from core damage -- from the PRA. That information is
combined with a series of deterministic questions that cover areas for
which go outside the bounds of the PRA on shutdown safety, can it fail a
risk-significant component, can it cause an initiating event. Other
questions that may be outside the scope of the PRA.
And you blend those things together, those two pieces of
information, so the figure of merit really is a conglomerate of all that
information there. But at the end of there, there is actually a
determination of the basis of why a component is important and it's
documented in what we call a risk significance basis document.
DR. WALLIS: Everything is related to safety in some way,
and you're not going to --
MR. GRANTOM: Everything is --
DR. WALLIS: Then you can cut out a lot of things.
MR. GRANTOM: What you find out --
DR. WALLIS: Everything is related to safety, even if it's
MR. GRANTOM: What you find out is there's a lot of
components out there that are there to help you maintain the system.
There's a lot of components out there that are used to facilitate an
operator's ability to walk around to measure a certain parameter for a
And they may not have a direct nexus to safety. They're
there for another reason. There are some components that are there
merely just to drain the system and once they're restored in their
position, they perform no function.
So you'll see a lot of those differences.
DR. WALLIS: No matter how miniscule and fine-detailed your
PRA, there are some things which would never appear in it at all.
MR. GRANTOM: Never, never appear.
DR. WALLIS: And those are what you would call
MR. GRANTOM: I would call them non-risk-significant.
DR. WALLIS: Well, significant -- you need some sort of
MR. ROSEN: Yes, and those levels were defined in the staff
MR. GRANTOM: Those are defined.
MR. ROSEN: -- in the staff review of our --
DR. WALLIS: A cutoff point, if it doesn't have more than a
MR. ROSEN: We call certain things not risk significant,
DR. WALLIS: This would seem far more rational than I think
the old scheme, which is anything important to safety, NRC says -- is
that the old regulation?
MR. GRANTOM: Yes.
MR. ROSEN: We have -- I couldn't comment on that, but we
have a threshold established.
DR. KRESS: I think it was if it was required to meet the
criteria in your safety analysis report.
DR. WALLIS: Which was not based on risk necessarily. There
might have been some risk in it somewhere.
DR. KRESS: It had some risk thinking in it.
DR. WALLIS: But it wasn't specific.
DR. BONACA: It's coming from analysis still, but a
deterministic analysis where you didn't have a full appreciation for the
likelihood that they would contribute more or less. You couldn't make a
DR. APOSTOLAKIS: When you use importance measures, do you
use them for both CDF and LERF or just CDF? In other words, you
calculate Fussel-Vesley and RAW for LERF?
MR. GRANTOM: Right. We can do both. We normally use just
CDF, but we do provide information on LERF, because part of the reason
is there is a matter of just practicality.
We've pretty well defined the components that are important
to LERF and we've set them aside. We find that all the other
components, for the most part, are important to CDF.
DR. APOSTOLAKIS: So you know already what's important to
MR. GRANTOM: We already know what they are and they've
already been treated. We already know that population. So I don't need
to do sensitivity studies on that, not nearly to the rigor that I wanted
to do for core damage frequency.
DR. APOSTOLAKIS: It would be nice to have one such study,
though, maybe from Research, see whether there is anything different.
In other words, find the two importance measures for both LERF and CDF
and look at the rankings.
I'm pretty sure what Rick is saying will be truth, but you
MR. GRANTOM: You might come out with something different.
DR. APOSTOLAKIS: So essentially, CDF-based evaluation, the
MR. GRANTOM: For the quantitative portion, yes.
DR. APOSTOLAKIS: For the quantitative.
MR. ROSEN: But not everything. Not all our decisions are
based on model components. We're talking about all the components in
the system, 28,000 in 22 systems.
DR. APOSTOLAKIS: Yes.
MR. ROSEN: And those are -- a lot of the components, the
expert panel that I chair has to make decisions without a PRA.
DR. APOSTOLAKIS: Sure.
MR. ROSEN: Because that component has been screened out of
the PRA and was never considered.
MR. GRANTOM: I do want to add that it's very important to
know that a component was not included in the PRA. There is value in
knowing that information.
DR. APOSTOLAKIS: See, that is a mystery to me. For some
reason, we've got several times that some components are not in the PRA
and we have to do something about them, and for me, if it's not in the
PRA, there is a reason why it's not.
MR. ROSEN: Of course.
DR. APOSTOLAKIS: It probably doesn't play any role.
MR. ROSEN: There were no scenarios, no sequences in which
it played a role.
DR. APOSTOLAKIS: So did your panel level, Steve, take a
component like that and put it in the safety-significant and
risk-significant category? Are there any examples of components like
MR. ROSEN: Of course. In my next slide.
DR. BONACA: Before we move to that, let me understand,
however. So this slide, you are trying to relate a current jargon, like
safety-related, to risk information.
MR. ROSEN: Mario, what I am trying to do here -- what we
are trying to do is to get to this bottom bullet. That not
risk-significant should not be -- should be not safety-related, things
that are not risk-significant ought not be safety-related.
DR. BONACA: I understand. Because safety-related has a
clear meaning in regulation today.
DR. APOSTOLAKIS: And risk-significant, you really mean CDF
and LERF significant.
MR. ROSEN: Right.
MR. GRANTOM: I'm going to say that risk-significant means
the combination of both the quantitative and the qualitative.
DR. APOSTOLAKIS: Yes. But you don't mean individual risk.
MR. GRANTOM: Right.
DR. APOSTOLAKIS: These measures. You are using the
MR. GRANTOM: Right.
DR. APOSTOLAKIS: For the quantitative.
MR. GRANTOM: Yes. And this is important to buy into this
fundamental concept here or we really don't -- we really don't allow
ourselves to step forward any further in risk-informing.
DR. WALLIS: The question is what's risk-significant depends
upon sort of the level of risk that the society accepts, because you
could have -- if someone says, no, we're only going to tolerate CDFs of
ten-to-the-minus-six or something, immediately other things become
risk-significant which weren't significant before. So the cutoff is a
variable maybe, depending upon society's risk acceptance.
MR. ROSEN: Right. And that societal risk acceptance was,
we presumed, embodied and built into our graded QA program through the
negotiated thresholds that we worked on with the staff that they
ultimately approved, something that falls below -- let's see -- it was a
risk achievement worth of two and a Fussel-Vesley less than .01 and --
it's an and -- less than two for risk achievement worth. That is not
MR. GRANTOM: It goes in from the quantitative portion as
being low and then you still have to address the deterministic issues
that I talked about before. And all of that's delineated in the SER for
the graded quality assurance program in the risk-significance evaluation
process, how that works.
MR. ROSEN: So what we came to talk about was not what's on
the prior slide and just to give you some gratuitous stuff that maybe
you'll think about, but here the next slide is a very complicated slide
that we would like to go over with you that shows you some of the real
results of doing this, and it's really two slides. It's this one and
As you look on the right-hand slide, it says see next slide,
see next slide.
This is an 11-by-17 foldout in our plant and, unfortunately,
we can't use that 11-by-17 foldout here. But I think you'll get a good
feeling from looking just at this first and then go with me to the
second page later.
What this shows is that the first circle on the left over
here, that we evaluated 22 systems with 28,267 TPNS. What's a TPNS?
It's a total plant numbering system. Each component has an individual
So here are 28,000 tagged components at the plant and 22
systems. Down on the bottom, it says note one, inclusive of -- it's an
alphabet soup here. Those are just system designators. AF stands for
auxiliary feed water and CC stands for component cooling water and so
So we analyzed many, many of the important risk-significant
and safety-related systems in the plant and what we found right away, of
course, is that safety-related systems, all the components in
safety-related systems are not safety-related.
I'll pause and say that again. Not all components in the
safety-related systems are safety-related. In fact, if you add them all
up, in the 28,000 that we analyzed, 54 percent are safety-related and 46
percent are not.
DR. KRESS: Now, a lot of the regulations on the books are
there to do other things, like, for example, cut down on worker
exposure, to cut down on releases that won't hurt anybody, but are
significant releases because they're approaching levels that might hurt
-- they won't be risk-significant, but because the containment is not
failed or something, but a lot of the regulations are designed to
prevent those kind of releases at a frequency that is acceptable.
I have a feeling that a lot of those what I would call other
regulatory objectives, other than CDF and LERF, won't show up on here as
safety-significant, as risk-significant components, whereas they might
actually be of regulatory significance.
MR. ROSEN: Well, regulatory significance aside, we have an
expert panel whose job it is to not take the results that come out of
the graded QA working group without using its own expertise and
experience to address exactly the kinds of things you're talking about.
In fact, the working group does that before it ever brings
to us -- we have a graded QA working group. It's one of many working
groups that report to the expert panel.
DR. KRESS: That might be a good function for an expert
MR. ROSEN: The working group brings us those concerns and
brings us those -- they have the PSA input, they have one of Rick's
people on the working group. But they bring us those additional
concerns, as well, and say we increased the risk rank of this because of
those kinds of things, so you should understand that, and then the
expert panel considers all that before we approve the final documents.
DR. KRESS: That's why I was -- you know, I was going back
to your previous slide where non-risk-significant should be equivalent
to non-safety-related. I don't think that's exactly true, because of
these other considerations.
MR. ROSEN: If you think the earlier slide includes the
expert panel deliberations, it's still true.
DR. KRESS: If you include that as -- the word
risk-significant, but they may not be risk-significant.
MR. ROSEN: It is the expert panel's job.
MR. GRANTOM: That's what I was trying to allude to, is that
risk significance, at this point, is after both quantitative and
qualitative information has been blended together and a judgment has
been passed over these components as to what they are. That's where the
risk categorization actually occurs, the final risk categorization
MR. ROSEN: So now let me take you past the second bullet
here, the second circle, and talk about what we found, some very, we
think, important things.
The first thing we found, looking at the top, is that of the
54 percent of the components that are safety-related, only ten percent,
1,446 are highly risk-significant, 15,000 components come out of that
block, 7,000 of them are not risk-significant, even though they are
safety-related, 7,000 are not risk-significant, even though they're
We label that a commercial issue, because we're treating
those 7,000 components with the full panoply of our QA controls.
Coming at this block the other way is also instructive,
13,000 non-safety-related components, of which 20 are highly
risk-significant, 344 are medium risk-significant. We label that a
safety issue, because it's possible that we have not been applying our
full controls to these.
DR. APOSTOLAKIS: So you're saying then that currently classified as
non-safety-related SSCs are, in fact, some of them are highly
MR. ROSEN: Let me tell you about what some of them are.
Positive displacement pump at South Texas is a non-safety-related
component. Its job is to cool the seals of the reactor coolant pumps to
prevent seal LOCA, alternate seal LOCA system.
MR. GRANTOM: And it's kind of strange that --
MR. ROSEN: Excuse me. And it's also powered by a diesel
generator, not one of the main diesel generators, but what we call a
technical support center diesel generator, which is a non-safety-related
DR. BONACA: Why were they non-safety-related? I'm
MR. GRANTOM: Its primary purpose, the positive displacement
pump, was to satisfy the hydro testing that's required. That's why it's
put in the plant. It also is part of the chemical and volume control
system and it can inject to the reactor coolant pump seals and it's
powered from a non-safety-related, on-site diesel generator source.
So its primary purpose, although it was to hydro the system,
the RCS, it also can be used for this alternate seal injection method
here, which buys time for station blackout scenarios. And as a result
of the initiating event it relates to plays a fairly significant role in
the risk arena.
DR. BONACA: What you're saying, originally, the design was
not recognized as a main function and, therefore, was not called
MR. GRANTOM: That's right.
DR. APOSTOLAKIS: Would this be an indicator that safety is
improved, Graham? You were asking for quantitative indicators earlier.
I mean, they are taking 364 components which are currently not
safety-related and moving them up. That would be an indicator, wouldn't
DR. WALLIS: Especially if you have a measure of the safety
MR. ROSEN: Well, we clearly know the safety impact of those
components, because we have --
DR. WALLIS: Yes, you do.
MR. ROSEN: -- we have the risk achievement worths.
DR. APOSTOLAKIS: What you don't have is the impact of
moving of the QA.
DR. KRESS: Yes, moving it up.
DR. APOSTOLAKIS: If you impose QA requirements.
MR. ROSEN: In fact, our plant staff, when we talk about
these things, say, but we're already doing all the important things, the
things we think are important for those components, because we knew
intuitively they were important. And we said, but, yeah, but there are
some things we'd like to be sure of, we'd like some independent
verification of these things, we'd like some of the tools that we
normally apply to, for example, aux feed water systems components, we'd
like to apply them in these areas.
DR. WALLIS: You could show that some increased attention in maintenance
had reduced the probability of failure of these previously identified
SSCs, then you could really show you've got a safety impact.
MR. GRANTOM: The technical support center diesel generator
is a vivid example. It used to be basically a small diesel generator
that sat on a skid that was open to the environment and you could throw
a rock and hit it, you know, from here to there, it was easy to do.
It's now enclosed in an air-conditioned building. We had
problems with the generator output breaker on it for years. It
chronically was a problem, where it wouldn't start and load. They
finally, because of the information that we had here about this is a
risk-significant component, finally have repaired the output breaker.
The reliability of the component is much better than it was.
As a matter of fact, at one time, it actually was
unavailable for the better part of a year because of the generator
output breaker component. But because of the heightened awareness of
this kind of a program, you quantitatively can prove its reliability and
availability is better.
DR. WALLIS: This wasn't something that you did because of
regulation, evaluated the risk-significance and then did the right thing
to reduce the risk.
MR. GRANTOM: Exactly.
DR. APOSTOLAKIS: And you have to speak to the microphone.
MR. SIEBER: You had about 11,000 components that you've
decided were worthy of downgrading the quality requirements and about
364 of high risk or medium risk components that you felt should be
Did you compare the end result of risk significance after
you made all these changes to see whether the total plant risk went up
MR. GRANTOM: With regard to the non-risk-significant
components, quantitatively --
MR. SIEBER: It should have been zero.
MR. GRANTOM: -- they're zero.
MR. SIEBER: Right. Okay.
MR. GRANTOM: They're not even modeled within the -- they're
not modeled within the PRA and the deterministic questions that are
asked, they ranked zero or extremely low even in those.
MR. SIEBER: Right.
MR. GRANTOM: So from that point of time, I can take the
aggregate and I can still say zero plus zero is pretty much a zero, from
an aggregate effect on that particular point.
I found out that more cases than not, they were a detriment
to safety because they caused us to remove safety systems from service
at times. So there was a real hit there in terms of availability of
some of the systems.
One of the other items that I will point to is that the
performance feedback that goes with this, where we collect plant
specific data, is going to be something that's going to be fed back into
this process in which we are going to be collecting the aggregate,
because as the question was earlier, I don't necessarily know -- I was
able to give an example on the technical support center diesel
generator, but these other non-safety-related components, I don't really
know exactly what the different changes in controls might be and how
that will work with its reliability.
So to answer your question, part of it I can discuss. We're
still waiting for feedback to really give a real good quantitative
MR. SIEBER: But there is some miniscule contribution to
risk from each of the low risk-significant items that you wanted to
MR. GRANTOM: From the lows, yes.
MR. SIEBER: And the question is, is that balanced by the
safety improvements that you got by upgrading.
MR. ROSEN: We think unquestionably. We really see it as a
see-saw with all these grains of sand having almost no weight, versus
two or three boulders on the other side, completely overwhelming.
MR. SIEBER: To me, that's significant, because when you
talk about risk-informing regulations, the idea is that you don't want
to degrade the safety posture of the plant in the process of
risk-informing the regulations and it would be good to have some kind of
feel like this if you've quantisized it instead of just having intuitive
feel, that, in fact, when you risk-inform a program, that you end up
MR. ROSEN: It's been our view --
MR. SIEBER: And perhaps cheaper.
MR. ROSEN: It's been our view all along that this would be
a safer plant when we're done with this and will be a more efficient
plant, as well.
DR. APOSTOLAKIS: I thought the question was, maybe it was
answered, but that the risk-significance of these thousands of
components has been determined on a component by component basis.
MR. ROSEN: Right.
DR. APOSTOLAKIS: Now you take 11,000 of them and you're
moving them to another category, there may be a cumulative impact, but,
again, I don't know how to calculate that, because we're dealing with a
particular issue here that is not quantifiable.
MR. ROSEN: We don't think a risk-significant component, if
its performance degrades significantly, would matter. It's already not
risk-significant. We would detect it, in any event, because we continue
our feedback loops and the condition reporting process on all these
components. We don't take anything out of our reporting system.
In fact, we require it. It's part of our operating quality
assurance program that we look at these things.
DR. APOSTOLAKIS: That's an important element, I think.
MR. ROSEN: We would not -- we don't feel we'll have
unintended consequences that go uncaught.
DR. BONACA: But you may have, at some point in the future,
some higher component failure rate because of the replacement you're
using, but it would take decades to really change your component failure
rates and understand them.
So we'll never be able to measure it.
MR. ROSEN: It would take decades and then we might
MR. BONACA: But what matters is that, again, the example of
a grain of sand. I mean, if it is irrelevant, you don't focus resources
on that. You focus resources on something else which is more relevant,
and I think that -- and the intelligence provided in the field and to
the plant regarding what is important and what is not, I think, is an
important contribution to safety.
MR. ROSEN: I think you make a very good point about the
intelligence provided. These results are embodied in what we call the
risk significance basis documents, which are then transferred, not all
the background information, but the heart of it, into our master
equipment database, which is the tool that our work control people and
maintenance people access on their computer screens every day.
So when the man enters a tagged TPNS number, he gets a
screen that tells him the risk significance of that component right
MR. GRANTOM: I think there is going to be a lot of
interesting information that comes out of this, too, because there are a
lot of competing effects, what's the value of a preventive maintenance
activity in terms of the reliability that it gains for a component or
And I think what we're going to find is basically it doesn't
gain very much for components that aren't worth very much. Obviously,
it's going to help for components that are worth more. So, I mean, it
becomes into a risk management perspective here in doing this.
But there are going to be some competing effects along that
line, which until we get the feedback and data and we start collecting
data -- obviously, a lot of these components or most all these
components are still embedded within the scope of the maintenance rule.
So there are controls that are there in place, both from a regulatory
point of view and also our own corrective action programs identify
non-conformances and deficiencies with these components, also.
So we're going to have to -- there is going to be a little
bit of a lag time, but I don't know that it would take decades, but I
think we'll be identifying component deficiencies, just like we have
been doing all along for both balance of plant side and NSSS sides.
MR. ROSEN: A few other insights that come from this graph,
chart. If you look at this circle, forget about these two for a minute,
if we had designed this plant starting today with the knowledge that we
have, we might not have a new set of regulations. We might only have
And this circle tells us only the risk significance. It
doesn't say safety-related or not safety-related. It tells us that
between low and not risk-significant.
Eighty-six percent of the components in these systems are
either not risk-significant or of low risk significance. Only this one
and this one, 14 percent together, the yellow and red, are the
components upon which we should focus our attention, from procurement,
design, construction, whatever point of view, operations point of view.
That's where the bang for the buck will be, from a regulatory
perspective and from an operator perspective.
I think that's an important insight and it's one, of course,
we didn't have when we built this plant.
So, now, let me ask you to make believe that this fits with
the other slide, this is an 11-by-17 graph. It looks something like
that. And pull over to the next slide and see what happens, what we do
with this information.
It's nice information, but the bottom line is what do you do
MR. GRANTOM: This goes to implementation here. If you can
imagine how do we actually take this information, this risk-significance
information, at this component level, and merge this to implementation
in plant programs, and you're looking at one program here, just one
program out of probably 200 that we've got at the station, this is one
program treatment here in the maintenance control, in the work control
MR. ROSEN: Right. This is our work control program, work
process control program. This is the process through which our work
control people tell the maintenance people what to do, how to do it.
And for the things that are on the top line, which you will
recall came across from the safety-related high risk significant
components, what they tell them to do is do what you've been doing.
Five percent of the total systems, structures and components population
continues to be restricted to the full maintenance application, continue
to do all those things, the big package, the fully detailed work step by
However, 94 percent of the population is now available,
leaving this piece out, for two other forms of maintenance, minor and
tool pouch maintenance, which are much more efficient ways of doing
maintenance from the point of view of, first, building the packages, the
work control packages, and from the point of view of actually
implementing that work in the field.
So what we see here in this chart is that a huge population
in the plant becomes available for more effective and a more streamlined
maintenance planning and maintenance work control.
MR. GRANTOM: I can add just a couple of little things here.
This is an important slide to realize from the implementation point of
what it actually means to the utility at this point in time.
When something is opened up for tool pouch maintenance or
minor maintenance, that means I can go work on a component, I can send a
craftsman out on a component, whereas before he had to have an entire
pre-planned package, with QA/QC controls, he had to have independent
verification, he has to have detailed work instructions on every step,
initial and signoff, and it may be a job that he's done 100 times and a
simple job, but when you open a component up to be able to do a tool
pouch maintenance or minor maintenance, he needs very little. Tool
pouch maintenance needs very little, it's almost an electronic
documentation of this is what -- I replaced the packing on the valve,
per my training, and it goes to skill of the craft.
We opened up more scope for the plant to be able to use
skill of the craft, and these are the areas where the plant is going to
grow into. Skill of the craft will improve as a result of this because
there is more of the station that they can apply it to.
Minor maintenance is a little bit more complicated. But in
the past, if it was safety-related, you didn't have the option to do
minor maintenance, you didn't have the option to do tool pouch
maintenance. You had to have full pre-planned package for every job
that was associated with a safety-related component.
And that's a lot of money involved and a lot of process and
a lot of burden involved in administering that. And you can carry it
all the way from pre-cal/post-cal of all the instruments to the actual
archiving of the records in the records management system that have to
be done for every one of those little jobs, but now with this type of
grading that's allowed, well, you can handle it tool pouch, which has a
much less restrictive and minor maintenance.
DR. APOSTOLAKIS: So the work processes are different.
MR. GRANTOM: It affects activities significantly and this
is part of the other thing that a lot of people missed, I think, and
initially people thought about procurement, but it goes to the
activities. It's how you treat that component, the strategy you use or
that you're allowed to use to go and address a deficiency with a
DR. APOSTOLAKIS: Do you do any on-line maintenance?
MR. GRANTOM: Absolutely. At South Texas, we --
MR. ROSEN: We do a tremendous amount of on-line maintenance
using this kind of information on a basic PRA to help us understand
configuration risk and because of the extent of our on-line maintenance,
when we get to outages, we have a much more limited scope than we used
to and South Texas now runs outages typically -- we hold the US PWR
record for the shortest refueling outage, 17.5 days, and we typically
plan the outages, sub-20 day outages.
DR. BONACA: How do you treat the medium risk significance?
Do you do it as basic maintenance?
MR. GRANTOM: It's allowed to be in basic maintenance. I
don't want to mislead you either. They always have the option, even if
it's on a non-risk-significant component, if the maintenance task is
just extremely complex or complicated, they can still do it as a
pre-planned package. Basically you have to look at it as the spectrum
of maintenance treatments that they're allowed on a given component.
MR. BONACA: I was curious, because I saw that that is
appearing in this slide. I mean, essentially you kept the high risk
significance, which you do as full maintenance, and the rest you do
really under basic.
MR. GRANTOM: Yes. They can do -- full maintenance is only
allowed on the high risk significance items. Basic maintenance is
allowed on the others, but you can always upgrade if you have to. You
could do full maintenance on a medium risk component or a low risk
component and that would be basically predicated on the complexity of
the task and that type of thing.
One thing I did want to bring -- I know this is talking
about maintenance application, but if you could think about it, if you
could use your imagination here, just to scratch out the word
maintenance and do design engineering or system engineering or
operations or training.
You name an organization --
MR. ROSEN: One of the key places where we've taken some
advantage of this, it doesn't show on this chart, other than
procurement, which we still have a problem with, is in preventive
maintenance frequencies. How often do you do preventive maintenance on
Well, someone ought to be taking into account the risk
significance of that component, as well as its performance. So if a
component has been doing well performance-wise, every time we take it
apart it's okay, and it's not risk-significant, surely we should be
extending the frequencies of maintenance, and we have. We've --
MR. BARTON: You've got the same option on a component
MR. ROSEN: We do.
MR. BARTON: For preventive maintenance.
MR. ROSEN: We do, but there we would go with more -- we
would be a little bit more concerned about doing that with a component
that has had a good performance history, but it's very risk-significant.
Maybe there we would say the reason it's had a good performance history
is because we're doing so much maintenance.
So let's keep on doing it there because of its risk
significance. On the other hand, for a component that's not
risk-significant that's had a good performance history, we're willing to
tolerate a little increased -- more failure in service, if we see it,
because it doesn't matter. It's a so what. When it happens, we would
say, okay, I guess we went too far with that, let's tighten the
maintenance back up again and without having incurred any additional
risk during that period.
MR. GRANTOM: I feel compelled to make one other quick
statement here. I know we have this word maintenance, but I'm going to
ask you also to use your imagination and think of motor-operated valves,
air-operated valves, whatever generic letter comes out that wants us to
do a treatment on all safety-related components, if they use that word
or that terminology with us, in this type of a process, what we would do
is we would apply this type of a process to figure out what's the proper
treatment for the high risk significant components versus the low risk
significant components, and that's why this is -- if you can get to that
point, you can see that this directly leads into option two and it also
directly leads into option three.
DR. APOSTOLAKIS: Has the staff, though, accepted your new
MR. ROSEN: Yes.
MR. GRANTOM: They've approved the SER that has the
description of this process in it.
DR. APOSTOLAKIS: The process.
MR. GRANTOM: Yes.
DR. APOSTOLAKIS: But have they actually accepted --
MR. ROSEN: The staff has not done an inspection of this,
except as the residents have looked at it.
DR. APOSTOLAKIS: According to the staff now, do you really
have only 3,400-and-some high and medium?
MR. ROSEN: We have the same number of safety-related
components. The non-safety-related components we always have. We
haven't changed that. What we've done is told ourselves what of those
components -- which of those components are risk significant. We have a
new set of documents called risk significance basis documents that go
alongside our design basis documents and we are following the process
that we told the staff we would follow in the OQAP when they approved
it, and they are welcome to come and look in great detail, down to the
last element of detail with us.
It's a fully scrutable set of processes.
DR. APOSTOLAKIS: So are you implementing what you're
showing right now?
MR. GRANTOM: Yes. This procedure right here that does
this, if you can see down here, it says this was incorporated in plant
general procedure 03ZA90, Rev. 20. This was incorporated in there.
MR. ROSEN: The answer is yes.
DR. APOSTOLAKIS: So the fact, then the staff has accepted
your classification here.
MR. GRANTOM: They've accepted the classification, but Mr.
Rosen points out the point, and you're where we are, I believe, is even
though we have this information, we haven't changed anything. All the
components are still safety-related and we couldn't change the
classification of them being safety-related to something else because of
the special treatment requirements and the other things that came into
play that we heard about earlier.
So we haven't really been able to achieve the benefits or
the full implementation because of where we are.
DR. APOSTOLAKIS: So you are still treating all 15,221
components of the previous slide the way everybody else is treating
MR. GRANTOM: Sure.
DR. APOSTOLAKIS: Even though you have done all this
MR. ROSEN: It's required by regulation.
MR. GRANTOM: We can do our maintenance different now,
though, because I didn't have -- we didn't have to go to the staff to be
able to do a different treatment on here and we do have a graded QA
process, also. So we're able to implement something on the activity
level, but I can't take the component and reclassify it to be a
MR. ROSEN: So how did -- you know, we tried to get at that
and I think we talk about that in the next slide.
DR. POWERS: Before you step forward. You have, at the
bottom of your slide, inclusive of a variety of systems, with a lot of
acronyms, most of which I can figure out. Would you point out to me
which of those are your fire protection, fire shutdown, safe shutdown,
fire suppression, fire detection system?
MR. ROSEN: Those are not in there.
MR. GRANTOM: I can real quickly go through there. There is
aux feed water, component cooling water.
DR. POWERS: I don't need a translation of them all. I want
to know which ones are --
MR. GRANTOM: Fire protection --
DR. POWERS: And if fire protection is not in there, does
that mean that in your risk assessment, you didn't include fire as
MR. GRANTOM: No. It means we haven't analyzed the system
MR. ROSEN: Fire protection equipment is still being treated
just the way we always treated it. We haven't applied a risk
significance basis to any of that equipment yet.
DR. POWERS: I guess what I'm trying to understand is how
complete -- when you go through and do your risk classification in high,
medium and low, how complete is the PRA? Does it include fire as
MR. ROSEN: Yes.
MR. GRANTOM: Yes.
DR. POWERS: And it includes operation during shutdown modes
and things like that.
MR. GRANTOM: It includes -- our PRA includes external
events, special interactions, fire PRA, seismic PRA.
MR. ROSEN: And we have a shutdown model.
MR. GRANTOM: And we have a separate shutdown model, but we
didn't use the shutdown model for this analysis because I had not
brought it into configuration control space yet. We have specific
requirements and procedures for our at-power model to be in
configuration control space and that was done qualitatively.
We will be bringing the shutdown model into configuration
control space and we'll be evaluating those components. But shutdown was
considered and is an important fact in the risk significance
These represent the systems that we have analyzed. We have
not analyzed all the systems at the plant. When we get to fire
protection, then it would be looked at in that light and you would get
the fire scenarios that come out of the PRA from that point of view.
Until the system is analyzed, there are no changes. It's
treated the same as it's always been.
DR. APOSTOLAKIS: What's your low power shutdown CDF
compared to ours? Do you remember?
MR. ROSEN: Yes, I do. We don't think it's as certain.
There's a lot more uncertainty in our low power and shutdown model than
we think is in our at-power model. Our at-power model for South Texas,
remember, it's a unique plant, with three safety trains. Right now
we're saying our core damage frequency for at-power is about
1E-to-the-minus six. For the last three outages at South Texas, the
core damage frequency that we're calculating is about
4E-to-the-minus-five. We think that's too high by a factor of two,
maybe a factor of four.
I said these same things at the lower power and shutdown
risk workshop that the NRC conducted and my conclusion was that although
our model for shutdown was as conservative as -- probably conservative
by as much as a factor of four.
If you grant that, then 4E-to-the-minus-five becomes
1E-to-the-minus-six, by dividing, and we think we are -- well, no, I'm
DR. APOSTOLAKIS: Why?
MR. ROSEN: We're talking about --
DR. APOSTOLAKIS: A factor of ten?
MR. ROSEN: I'm sorry. The 4E-to-the-minus-five is probably
more like 1E-to-the-minus-five.
DR. APOSTOLAKIS: Yes.
MR. ROSEN: And our model now is showing around 1E to 1.3 or
something like that, at-power model, one to 1.3, 1.5 perhaps,
So what we're saying is that shutdown risk is very
significant at South Texas.
DR. APOSTOLAKIS: It's an order of magnitude, roughly.
MR. ROSEN: Well, not quite so much, but at least as -- on
the order of being important and the reason I'm hedging is because we're
not so confident of the shutdown model and its directly comparability to
the at-power model. But the conclusion we were able to reach, without
any big debate at South Texas, was -- what we could agree on was that
these risks are comparable.
The amount of risk we get in an outage, one outage, is about
equal to the amount of risk we get throughout the year from the at-power
DR. SEALE: That's very interesting.
DR. POWERS: Can you give us an insight on what your fire
MR. GRANTOM: It runs at about close to 5E-minus-seven for
MR. ROSEN: Part of the reason for that very low number, Dr.
Powers, is that South Texas is a heavily segregated plant, with three
MR. GRANTOM: Yes. Very separated systems and we find that
most of the shutdown risk occurs in the one window of the front-end
mid-loops for a PWR, and that's where you generate the most risk.
MR. ROSEN: One of the insights about shutdown risk that we
have is that it's very different than character from at-power risk.
At-power risk tends to be relatively homogeneous. Shutdown risk is very
heterogeneous. It comes in very, very distinct peaks and clumps. There
are periods when it's very safe to be shutdown, when you have the core
offloaded, for example, and there are periods when it's not -- when you
have some higher risk, for instance, at an early mid-loop operation.
MR. MARKLEY: But you don't do a 17-day outage offloading
MR. GRANTOM: We offload the core completely.
MR. ROSEN: We've always done that.
MR. GRANTOM: Full core offload. What you don't see is all
the train-related work that can now be done at power, local leakage rate
test that Appendix J requires us to do that are not risk-significant,
that we --
DR. APOSTOLAKIS: Did you go, in your low power shutdown
PRA, beyond what the two national laboratories have done for Surry, I
believe, and Grand Gulf, or did you follow more or less the same
MR. GRANTOM: I don't understand the question.
DR. APOSTOLAKIS: A number of years ago, Sandia and BNL
published PRAs for shutdown modes that were kind of selective. They
selected the modes and they said in our judgment, these are the most
important ones, and they did the PRA.
And human error, of course, was very important. I mean, did
you do a similar thing in your PRA or did you go start it from scratch
and you identified all the modes of operation and initiating events and
so on? How complete is it compared to the power PRA?
MR. GRANTOM: We were able to use most of the systems
analysis directly from the at-power portion, and adding just a few
systems, like spent fuel pool cooling. We did some initiating event
screening and reviews, but it was not as rigorous as the at-power model.
Plus, we didn't go through all of the plant modes and do
specific plant initiating event searches during each plant mode for
power descension and then back up for power ascension.
Primarily looked at standard initiating events that had
occurred at that point, and that's why I say we're still moving that
into configuration control space and still improving that model.
The systems analysis portion is very good. Some of the
phenomenology I think is still rough and I also think that there is more
than can be done in the initiating event screening.
DR. APOSTOLAKIS: How about the human performance?
MR. GRANTOM: Human performance is in there for operator
actions that are included in -- currently included in our procedures and
those are included. So in that regard, I think we have a fairly good
treatment of human actions.
If you're going to expand human actions to be human-induced
actions, such as drain-down events or other types of things, that's a
little bit looser, rougher.
MR. ROSEN: That's why we have that factor of four feeling
that we're not so sure about the exact number of shutdowns.
DR. POWERS: Your factor of four was going down and it seems
to me that this looseness takes it the other direction.
MR. ROSEN: Potentially. But in discussing with the staff,
most of the people think it's conservative overall.
MR. GRANTOM: We do have some conservative treatment of the
initial cool-down period and going to operator actions and losses of
coolant during the initial cool-down. We have some conservatism there
that we're trying to work through. We have a different result than
other plants have, for some reason.
MR. ROSEN: Okay. I'd like to turn to the final few slides,
and leave a few more gratuitous points with you as you consider going
It seems to us that as you go forward to risk-inform the
individual sections of Part 50, the lesson learned from our experience
was that we really can't get the procurement benefit that we needed and
that we intended to get and justify the endeavor on the basis of without
the special treatment requirements being addressed.
We tried to get at this problem -- next slide. We tried to
get at this problem just recently with the staff. We had a meeting with
Gary Holahan and his crew, when we described this problem to them, and
the judgment we got was that we could perhaps take our not
risk-significant components that are safety-related and change the
classification to non-safety-related from safety-related on the basis of
-- and you do that using 50.59.
And that was a bit of a revelation to us that the staff felt
that way. We weren't so sure that that was okay. So we wrote a letter
to the staff and posed that question, with lots of communication back
and forth to the staff.
And we got a letter back from them on May 28th, saying, yes,
indeed, you can change the classification of safety-related components
from safety-related to not safety-related if they're not risk
significant and you can do that using Part 50.59, which means that we
don't have to come in on a one by one basis and ask permission.
And I was kind of elated when I got the letter and I was
sort of surprised by the staff's reaction, my staff's reaction to it and
the licensing staff's reaction, and the design engineering staff's
reaction, which was it's ambiguous.
If you read the letter, the cover letter, the first two
pages before the signature, it didn't seem very ambiguous to me.
When you read the nine concerns in the enclosure attached to
it, taken together, it comes out ambiguous.
So the fact of the matter is, the bottom line is we haven't
done anything, because we're still on square zero trying to figure out
what that letter meant and did it really give us the room to maneuver.
MR. GRANTOM: I'd like to just add one quick thing to what
Steve said. The good analogy that I guess our design engineering
manager says, he thought it was kind of a yellow light that was turning
red and we thought it was a yellow light that was green but had gone to
So we weren't really -- and one of the discussions that came
up was, well, we need to just not use safety-related, let's just leave
safety-related and let's just call it risk-significant and just treat
things differently, risk-significant, and just leave it safety-related,
because we're going to get in too much trouble if we try to take
something that's safety-related and move it to non-safety-related.
So we're kind of stuck in this quandary with the design
DR. WALLIS: It seems to me it's a pretty straightforward
question. You want to take these from this box and put them in that
MR. ROSEN: And we want to use 50.59.
MR. GRANTOM: Because when you get to the 50.59 question
that says, well, the change may cause an increase in the probability of
a malfunction of a component, "may" is a pretty open-ended thing.
And they hung up and that defaults to USQ at that point in
time and staff approval and --
MR. ROSEN: Plant operations review committee approval,
nuclear safety review board approval, submission to the NRC.
MR. GRANTOM: Everything locks down, the whole application.
MR. ROSEN: We get to unreviewed safety questions, you stop.
DR. POWERS: Do the revisions that the staff has put forward
and proposed for 50.59 get rid of that problem for you?
MR. ROSEN: I don't know yet, Dr. Powers. We haven't really
finished our review of that. Our initial discussions with our licensing
people looking at the current proposed revisions, and I guess NEI will
have their own view of it, is they don't go far enough.
MR. GRANTOM: I don't believe it goes far enough.
DR. WALLIS: I just wonder if these are minimal changes or
MR. GRANTOM: It doesn't allow quantification, it doesn't
allow the de minimus treatment, and, to me, it's not that risk-informed.
DR. KRESS: If the new risk-informed 50.59 came out, you
would recommend that some measure of RAW, which is what you're probably
-- and these other ways, would be a good way to decide 50.59 space,
MR. GRANTOM: We would probably say that the
risk-significance determination process that has been reviewed and
approved by the staff is an appropriate way to answer the 50.59 issues
there and that we could go forward. We don't even think we would have
needed the clarification letter or the exemption request we're dealing
DR. KRESS: I'm not quite sure your cutoff value for the
line between high, medium and low is, in terms of risk significance, the
right one to use for 50.59. But I certainly think that's an attractive
thought to look at that.
You may want to -- for 50.59, that number may want to be
MR. GRANTOM: Keep in mind, we wouldn't just use the PRA, we
use the blended process.
DR. KRESS: Blended process.
MR. GRANTOM: At the end of that. So it would still be a
blended process that we would -- and we probably categorize things as
non-risk-significant components, the low risk significant components
would be candidates for this reclassification.
MR. ROSEN: In any event, however it's done, this bullet on
this slide, that to be fully effective, Part 50.59 needs to
unambiguously allow risk-informed judgments to be applied, is the
DR. APOSTOLAKIS: So the statement of minimal changes is
ambiguous, in your judgment.
MR. GRANTOM: The same as "may cause," the same as "may
cause," same as "could cause."
DR. WALLIS: Thank you very much.
DR. APOSTOLAKIS: Are you dismissing them?
DR. WALLIS: No. That's what I have been trying to say,
DR. BONACA: It doesn't allow you to make your calculation
DR. APOSTOLAKIS: Do you want to continue?
MR. ROSEN: Yes. We have only a few other points we wanted
to make, and that -- whatever is done to Part 50.59, it must allow the
programs, activities, structures, systems, and components to be graded
commensurate with their importance to safety.
That's what we have been trying to do here and have found it
very difficult in the current environment and are only getting part of
the benefit that we anticipate. Of the $2 million I hope to get per
year, we hope to get per year, we're probably getting ten to 15 percent
of it out of some of the other things we're doing, that we could do
within our authority, without crossing the boundary.
DR. WALLIS: The NRC policy only allows ten percent at a
MR. ROSEN: Well, it's a credit I think to our thick hide
and our perseverance that we're still continuing, nevertheless, even
though we haven't gotten the benefit that we expected. We think it's
the right thing to do for the industry and South Texas to do this thing.
MR. GRANTOM: I don't know exactly how it can be done, but
I'm certain there are ways it could be, but you think about the special
treatment requirements and you would think, in a reg guide or in some
other guidance document, that there would be a strategy for high risk
significant components, that there would be a different strategy for
things that you would consider to be low risk components.
That's not unlike the approach that's being used in the ASME
arena for the O&M code cases that are coming up for motor-operated
valves and pumps. They're using a similar kind of approach there.
And when you get to option three and you're going to go and
look at the individual regulations themselves and change their technical
content, the regulations, at least it's my feeling, need to have the
flexibility in there to allow a grading to be done, a different strategy
commensurate with its risk significance, and to address the ones we were
talking about there from the second bullet, or have it handled in a
And I think at that point in time, option three becomes a
very tractable thing. You have the scoping taken care of, you have the
risk significance process that puts them in the right classifications,
then you have the right strategies per the regulations that build the
risk information into it. So I think it can work very well.
MR. ROSEN: Our final point is only this last one on the
slide. That we are preparing an exemption request, so that we can move
ahead and get the benefits that we anticipated. We think that is
directly in line with what option two envisioned and we're moving in
that direction, and Rick is holding it in his hand. The only reason
he's not dropping it on the table is it hasn't been signed by our
executives yet, but I expect it will be shortly.
MR. GRANTOM: Maybe even today, but I was pushing for at
least Monday to have it signed.
MR. ROSEN: So with that, thank you very much.
MR. GRANTOM: Thank you very much for your time.
DR. APOSTOLAKIS: Thank you very much. This was very
informative, again, as it has been in the past.
Mr. Heymer, you have something to say.
MR. HEYMER: Very briefly.
MR. ROSEN: We'll stay here in case there are any questions.
MR. HEYMER: Good afternoon. My name is Adrian Heymer. I'm
a Project Manager at NEI in the regulatory reform group.
I apologize for not having my name on the slide. I swung
past this morning and picked them up and it wasn't until I got on the
train that I realized it hadn't done the formalities.
Steve Floyd, our Director, sends his apologies. He had a
conflicting engagement. Otherwise, he would be here speaking instead of
I think what I've heard here today is gratifying. The
industry is pleased to see that the NRC has decided to move forward in
improving its regulations, and we see that it really focuses on
improving the efficiency and the effectiveness of the regulations using
risk insights or blended approach of risk insights, operating
experience, and deterministic analysis.
The NRC has moved forward as we go into a restructured
environment, just not in this area, but in other areas. They have been
very proactive and we're pleased to see that.
I think, to be frank, we would like to see a little bit more
movement, especially on option two than we've seen to date. But I was
gratified to see some of the points that were made today and I think
some of our thoughts are beginning to converge, especially when I saw
some of the criteria that were out there.
What I want to talk to you today is a little bit about our
overall implementation plan we see for option two and three and we have
a slightly different twist on things, but hopefully that will improve
We, in fact, believe there are four phases, we call them
phases, perhaps that's the wrong term, because they're not really
sequential. They do run in parallel and we do think it leads off the
oversight and the new NRC licensee assessment process, the oversight and
enforcement activities that are ongoing. There's nine pilots underway,
as you well know.
And to lead off into the other areas, and we think the other areas are
the SSC scope, which is what we see as option two, and we think that is
just, as the NRC explained, coming up with a definition and we think the
definition is associated with safety significant.
We have tried -- we have a small task force at NEI, along with a working
group, the regulatory -- the risk-informed regulation working group and
the Part 50 task force, floated a couple of suggestions for such a
definition associated with safety significant SSCs. We're still working
on that. It's not an easy task, but we hope that at the next meeting
with the NRC staff to at least get down and start discussing what
perhaps that proposal might be for a definition.
And I don't think it's too far removed from what I heard
today in some of the slides.
Phase three is the technical requirements. We think this is
a longer-term action plan. We think we can get on the move on the phase
two, option two pretty quickly.
On phase three, I think you've got to do it with a little
bit more thought than just say do it. There are some issues that are on
the table there, which ones to go for first, where is the most benefit,
is there consistency. We've been really focusing on phase two. We have
given some thought to phase three, some of the regulations jump off the
page at you fairly quickly, and at first sort of gut feel is let's go
after that, but I think both the NRC and ourselves are smart enough to
say let's wait and see.
DR. WALLIS: Can you give an example of the ones that jump
off the page at you?
MR. HEYMER: 50.46.
DR. WALLIS: What part of that particularly?
MR. HEYMER: I think when you look at that and you look at
the risk-informing of the regulations and what is covered by ECCS and
then look at the way we treat some of those activities and then see what
we've come up with in some of the risk-informed pilot projects that
we've worked on, I think that one says perhaps it's worthwhile at least
looking at that or taking a look to say is it worthwhile assessing that
from a benefit perspective first.
DR. WALLIS: I was wondering if they jumped off the page
because you already saw there were benefits. You haven't gotten that
MR. HEYMER: I think deep down people say yes, there are
benefits. But if you then follow through and say, well, quantify those
benefits, are you sure the benefits are there, how does it relate, if
you really changed that regulation, really think it through, I think
you've got to stop yourself and do that before you actually plow into
that. It will be a fairly intense exercise.
Phase four is the administrative regulations and we pulled
this out separately because we think there are some activities that are
underway associated with 52.73, 50.72 and 50.73, with 50.59.
That if you go forward now with a rulemaking, you've got a
rulemaking on top of a rulemaking and we're not quite sure how that
would turn out.
You've got a rulemaking exercise and then we come in with a
petition or the NRC goes forward with another rulemaking, you've got two
going forward, and we think that might be a little bit complex and
difficult to handle. So that's why we pulled those out separately.
But we see those going on as parallel activities. Going
back, phase one is off and running. We would like to -- we are
considering moving forward and coming in with a package associated with
phase two and we've been given a target date to look at of September the
30th, and I will speak more a little bit about that in a moment or two.
Suffice it to say we got rulemakings, we've got guideline
developments, as they're needed. We've got pilot projects as they're
needed and as necessary. And I think we would agree with some of the
comments made here today that you don't need the pilot projects for
everything, but there are some activities that it's worthwhile having a
pilot project for to test implementation phase and to test the -- assess
the benefits, are the benefits really there.
DR. WALLIS: Have you found that this South Texas project is
useful for developing your arguments for phase two?
MR. HEYMER: South Texas is part of the graded QA activities
that went on before. They're very much seen as an industry icebreaker.
I think the industry -- there is a lot of activities that South Texas do
that has its benefits on an industry-wide basis. That is not to say
that everybody will go down the same path as South Texas or go to the
same extent as South Texas has done in some of their analysis.
I think South Texas has gone the extra mile to prove the
process, but perhaps we don't have to go there.
DR. WALLIS: Because they seemed to get to a point where
they thought they had proved a lot of things and then the NRC came back
with something called ambiguous. I don't think you want to make a lot
more out of that.
MR. HEYMER: Well, you've got to change the regulations, but
I think what we would like to get to a position is where the industry
gets as far as perhaps South Texas did and because of what the work
South Texas has done, gets an acknowledgement and an understanding that
that is satisfactory, rather than having to go and perform further
DR. APOSTOLAKIS: Is that what you said?
DR. WALLIS: I thought you'd be saying this is great support
of our position. Something needs to be come about SSC scope and look at
the South Texas analysis and look how NRC didn't respond to it in an
I'm surprised you're not making it as a strong argument for
moving things along.
MR. HEYMER: I think what we would like -- what we intend to
do is to come in and come up with a process that's very similar to South
Texas, but as the South Texas people here have noted, that they've come
up and running to an area where they're not quite sure if they've got
the go-ahead of not. And I think what they've done, many in the
industry believe is sufficient and I think if they have to do much more
and if the industry has to do a lot more to actually go and get the full
measure, there may be pause for thought.
DR. APOSTOLAKIS: I guess the question is, or at least the
way I understand it is NEI using the lessons learned from the South
Texas experience. For example, in option three, it seems to me the
message is clear from them that you should risk-inform 50.59, at least
one of the messages.
MR. HEYMER: And we totally agree with that. South Texas is
on our task force. They're helping us write the guideline for an
industry-wide guideline for categorizing SSCs into safety-significance.
They're totally immersed in the other risk-informed activities that we
So I think the lessons learned are folded back in and
perhaps tempered a little bit by some people saying, well, perhaps if we
did it this way, it might be a little bit better, and South Texas can
bring their insights into that with their interactions with the NRC.
DR. APOSTOLAKIS: Okay.
MR. HEYMER: One other element that is not up there, which
is just as important, and I think the NRC staff and the industry really
focused on it, was a communication plan for each of these four phases to
get the message out.
The first bullet we've really touched on, a model improving
safety, we really said, is improving the focus on safety in order that
we can have a much more efficient and effective method of allocating
resources to those things that are important.
There has been a fair bit of discussion today about
voluntary and selective. I think if you look at the composition of the
industries today, where we have four NSSS vendors with various designs,
we have a multitude of architect engineers, consultants and contractors
who help us improve and implement those designs. We have 60-plus
owners, each with a different view of where they want to go.
There is a recognition that we are safe today. If we're
going to go forward, I want to make further improvements in efficiency
and effectiveness, and I think that should be on a voluntary basis.
I think once you've made that step and say, yes, I want to
go down through option B to Part 50, which is risk-informing the
risk-informed elements of Part 50, I think once you start getting into
that, to start picking and choosing, you might run into some efficiency
problems of your own.
So it might be self-revealing that once you start down the
path, you keep going. How far you go is a matter of resources to
implement those activities and the benefits that they have, but I think
once you start going down that path, you are faced with do I have two
different systems, do I have two different inspection regimes to deal
with. It does become a little bit more difficult, but I think the
option should still be there.
We think the phased approach to rulemaking and what we see is that
probably would be the definition and introduction of the definition of
safety-significant, the introduction of option B, and the categorization
of SSCs into two new buckets or groups would be one rule, and then be
applied individually to various regulations.
That's on the scoping side. On the technical side, it will
be a number of rulemakings for each specific regulation. So on one
hand, you have one rulemaking and a phased implementation associated
with each regulation based on benefit and need. On the other hand, you
have numerous rulemakings associated with specific regulations that
would actually make them risk-informed.
So rulemaking one, which is the scope, would just replace
the term safety-related with safety-significant, in simple terms.
There's a little bit more to it than that.
Introduce the term safety-significant SSCs and introduce the option B,
on the technical aspects you would go into and risk-inform, for example,
50.59 under option three.
As we move through this, we think the scope and all these
activities, there needs to be a consistency and what we're looking at is
using just not PRA itself, but a blend, a blend of risk insights from
the PRA, operating experience, new analytical techniques or new
technologies, as they've come along in the past 20 or 30 years, that we
can take benefit of.
On the consistency issue, we've put this slide up a couple
of times to try and show what we mean. It's that if you look at phases
one and two, the outside oval or egg is the existing regulatory scope,
as we see it, and we have a much more focused approach, looking at the
things that are really are important from the assessment perspective.
It's only right that we focus on the regulations and the
oversight process focuses on the same things. But when you look at the
regulations and you go to a risk-informed, it's a little bit more than
just this oval here, in fact, because you are using risk insights,
perhaps you're not talking about -- it's more perhaps -- you're not
talking about just being corralled by the design basis events. You're
going beyond that.
So, therefore, the more focused circle moves somewhat out of
the regulatory scope and it's only right that once you start doing that,
that the oversight process also does that. So that's how we're trying
to ensure the consistency and why we believe it's important that we move
ahead and adjust the SSC scope and to make phases one and two, as we
called them, consistent.
DR. APOSTOLAKIS: So you're introducing now a new
terminology, safety-focused, scope and oversight process. I haven't
seen that before.
MR. HEYMER: I wouldn't say it's a new -- it's just a --
it's a way of summarizing what we're trying to do, is we're trying to
improve the safety focus of the regulations.
DR. WALLIS: I'm getting the impression that the shaded area
is smaller than the other one.
DR. APOSTOLAKIS: They're expected to be smaller.
MR. HEYMER: You're saying that the actual improved scope
could be bigger than the existing scope.
DR. WALLIS: You gain something on the left.
MR. HEYMER: Yes, it does. It's just a diagram.
DR. WALLIS: I don't see what it's got to do with
MR. HEYMER: What we're trying to show here is that the --
if you're focusing on a certain set of attributes, they should be
consistent with the attributes reflected in the regulations.
DR. WALLIS: The two ellipses should overlay.
MR. HEYMER: Yes.
DR. WALLIS: You take one above and overlay them.
MR. HEYMER: Yes.
DR. APOSTOLAKIS: So what would be the current --
description of the current system using this terminology? It's not
focused or is it overly focused?
MR. HEYMER: It's focused. What we want to do is improve
DR. APOSTOLAKIS: Improve the focus.
MR. HEYMER: These are some of the regulations that we're
thinking of looking at from the complete perspective. This is phases
two and three and also phase four. They're not all on here. Some might
call this a sunburst, some might call it a prickly pear.
DR. APOSTOLAKIS: Do you have any ideas how you would
risk-inform license renewal?
MR. HEYMER: I think it goes along the lines of the SSC
scope, looking at that, and then saying you -- that's just from phase
two. On phase three, I think if you're changing the regulations on
phase three, it's automatically picked up in Part 54. But I think some
of the wording in Part 54, I think you might have to adjust a little bit
to reduce the ambiguity of what are we really looking at here.
I think it's a fairly straightforward and simple change.
We've had a brief discussion with those that NEI who deal with license
renewal and we've put this up as a strawman to them. We are looking at
it. We may or may not go ahead with it, but at least it's worthwhile
putting on the table to make sure people recognize that we are looking
beyond Part 50.
One area that is a little bit difficult and we think will
require special treatment is Part 21, and that's just the way Part 21 is
written and relates to the term basic component and where basic
component is defined.
DR. APOSTOLAKIS: Would you remind us what Part 21 is?
MR. HEYMER: It's reporting of defects and deficiencies.
DR. APOSTOLAKIS: Okay.
MR. HEYMER: And whether or not we want to change that to be
associated with just safety-significant SSCs, but I think we then have
to take a close look at the Atomic Energy Act, because it sort of all
emanates out of there. So that is a little bit of a unique case.
DR. BONACA: For license renewal, however, I mean, you
wouldn't make a special exception for license renewal to change your
definition of systems and components which are safety-related. You
would have to change it under the current license term before you move
to license renewal.
MR. HEYMER: Exactly. But I mean, just from a consistency
perspective, if you're changing in Part 50, just make the same change in
Part 54 what you're doing.
DR. BONACA: It would be an extension, of course.
MR. HEYMER: Yes. That's all it is. It's a consistency
As regards categorization, this just sort of is a graphic
that depicts what we're trying to do. On the left-hand side here is
where we are today with the various terms. You go through a process by
which it's blended using very similar to what South Texas explained to
you, PRA insights, operating experience, engineering analysis, come out
of that, have an expert panel or plant review by knowledgeable people
and come up with two new or three new or four new categories.
We've put two down here. That doesn't preclude the option
that you might want to grade further in either major boxes. The reason
why I've got a dotted line there across here, you start off with this is
where we are today with the terminology and you just come up with one
safety-significant SSC and others you could have high safety
significance and low safety significance, but judging by the discussions
we had about four years ago about using such terms, we just came up with
safety significant and other.
But when you do this, obviously, some components are
safety-related will end up as other, and others that are
non-safety-related, as you heard earlier today, end up being in this
The thing is that when some of the components up here come
down here, they may be specifically related to a regulation and until
you go in and change that regulation, risk-inform that regulation, you
are still loaded with the regulatory requirements associated with that
system or component or structure.
DR. APOSTOLAKIS: What is the significance in the left
column there of having the dotted line and then important to safety,
significant to safety? Are these terms that are used in addition to
MR. HEYMER: Yes. In regulations, you have safety-related,
you have important to safety, you have important to safe shutdown.
DR. APOSTOLAKIS: And these are treated as if they were
MR. HEYMER: Semi safety-related, yes. You have quality
requirements, not necessarily the full Appendix B, but certainly
specific requirements laid on these. So that was just a pictorial
representation.-- using the same terms, but I think because we have so
much familiarity with it.
DR. BONACA: Could you put it back? Just for clarification.
I mean, that's true that the regulation had that classification there,
but South Texas originally put in these components into that because
they did not take advantage of the QA, and now you're going to go to QA.
What I mean is that Appendix B always allowed you to grade
the ranking or you didn't have a tool to do that before, now you do. So
your main obstacle now is 50.59.
So I'm trying to understand how you go from --
MR. HEYMER: Yes. I think 50.59 is one. I think culture is
another. I mean, you could probably do this by having the same -- using
the same terms. But I think because we have so much familiarity with
DR. BONACA: You're not changing Appendix B.
MR. HEYMER: No.
DR. BONACA: You're just simply modifying your rankings.
MR. ROSEN: The main obstacles are two, 50.59 and the
MR. HEYMER: This is a slide that when we discussed with the
staff, we spent quite a bit of time on it. The categorization is what
we just discussed over here on the left and as you go through that
process, you end up with basically two bins, safety-significant SSCs and
As you come across, you can grade here, so this box at the
top may be related to the EQ components. The whole box here maybe
related to the maintenance rule scope. And in addition to that, once
you've done the SSC scope, if you go in and look at the specific
regulation associated with that and risk-inform that regulation, it may
So the regulations really relate to the safety-significant SSCs. When
you drop below the line, except in the -- the exception that I just
described about safety-related SSCs that now become other or low
safety-significant, but still are linked to some specific regulation,
there is -- the regulations do not apply.
However, normal business practices would still be employed
and would still govern, because I think as you're well aware today, to
maintain a good capacity factor, maintain operation, maintain good
material condition, which is the only way you can do some of the things
we do today, you have to have reasonable and good controls, from a
Those controls will be down here. For SSCs that move up,
you would look at those and say, well, what's really caused them to move
up, what is the specific attribute that has caused them to become
safety-significant, and then you would apply controls that are
commensurate with those attributes or whatever that may be, whether it's
seismic, whether it's pump flow or whether it's EQ, what scenario is it
Oversight and regulation, here we say it's just the process
that's being put in place today. The performance indicator and
significance determination process. So on safety-significant SSCs, you
could get Nov or non-cited violations.
You could still -- there is still NRC oversight on the lower
category, but on event-driven basis, and if something happened down here
in one of these that resulted in the event, obviously you would want to
know why, from a licensee perspective, and I think that would just give
you a notch to go back and look at the categorization, but also we fully
anticipate and expect the NRC to be interested in that, what happened,
was it related to some change in categorization, what controls have you
put on that, what are you doing about trying to ensure that that doesn't
That's where we see ourselves going. In timeframe-wise, I
told you about phase two and we hope to come in in the fall. Phase
three, we do see that as a longer-term effort. We haven't really
focused on that. We're just beginning to think about what would be the
criteria that we would lay down to identify what regulations we would go
after first as regards risk-informing them.
And the administrative requirements, such as 50.72 and
50.73, once we see how those rulemakings are shaking out, we'll make a
decision whether or not to include them in the SSC scope or just break
them out and have separate rulemakings associated with those, perhaps as
On 50.59, we see that being linked, split, and we think some
part of that is associated with phase two at the moment, but the real
risk-informing of 50.59 would be an option three, phase three issue.
DR. APOSTOLAKIS: Is that it? Any questions?
MR. ROSEN: George, I would like an opportunity to correct
something I said earlier, just finally. Dr. Powers asked about the
comparison between low power risk and at-power risk, and I proceeded to
murder the answer.
The reason I did is I got the at-power model risk number
wrong, just from memory. It's between one and one and a half
DR. APOSTOLAKIS: So they're comparable.
MR. ROSEN: Shutdown is about 4E-to-the-minus-five, at-power
is between one and one and a half E-to-the-minus-five. And we think the
shutdown is too high by as much as a factor of four. So they are very
comparable and I could not imagine why I couldn't get that right, and
then I realized, after thinking about it, I got it off by an order of
magnitude. So appreciate the opportunity to correct that.
DR. APOSTOLAKIS: Thank you. Comments or questions from the
members? Well, we appreciate very much your coming here. It was very
useful, as usual. Thank you.
MR. GRANTOM: Thank you for having us.
MR. ROSEN: Thank you.
DR. SEALE: It's always good to hear from South Texas.
DR. APOSTOLAKIS: Do you want to go around the table now and
express some views or do you want to take a three minute break first or
Are you going to give me advice how to brief the committee
DR. UHRIG: It's the Chairman's prerogative.
DR. APOSTOLAKIS: I chaired this afternoon. My co-chair
should brief me. Do you want to go quickly around, give me some views?
I mean, are you happy, are you unhappy? Are you in the middle category,
medium happy? I guess the enthusiasm is overwhelming.
I don't know. Do we want to find a+ separate meeting date
with the staff or we do that off-line?
MR. HOLAHAN: I thought you were going to let us vote on
whether we were happy or not.
DR. APOSTOLAKIS: I have to use objective measures. I guess
no volunteers. So on that happy note, we adjourn. Thank you very much.
[Whereupon, at 5:31 p.m., the meeting was concluded.]
Page Last Reviewed/Updated Tuesday, July 12, 2016